Heat Exposure and Transit Use:
Travel Behavior and Infrastructure
A Collaborative Research Project between Arizona State
University’s Urban Infrastructure Anatomy & Sustainable
Development and Construction Materials & Methods Classes
Spring 2017
Authors:
Arishi, Jake / Azarabadi, Habib / Castaneda, Joseline / De Leon, Rio /
Dwyer, Kevin / Dzyuban, Yuliya / Fenais, Amr / Fernandes, Nagelle /
Gonzalez, Andres / Hoehne, Chris / Mueller, Joseph / Podduturi, Suhas /
Ramatheko, Nokwanda / Rodriguez, Jordan / Rogers, Joshua /
Rogers, Samuel / Sanchez, Christopher / Santana, Kristian /
Sawyer, Madeline / Sherman, Rachael / Venne, Hunter / Wills, Andrew /
Xie, Jiang / Yellasiri, Sai Rutwik Reddy / Zhang, Pengfei

Advising Faculty: Mikhail Chester
Editor: Mohamed ElZomor

Report No. ASU-SSEBE-CESEM-2017-CPR-001
29 May 2017
https://repository.asu.edu/items/43327

TABLE OF CONTENT
PROJECT OVERVIEW ...................................................................................................... 5
1

CHAPTER I: TRAVEL BEHAVIOR................................................................................. 6
1.1 INTRODUCTION..................................................................................................... 6
1.1.1 Problem Statement and Research Questions:.......................................... 8
1.1.2 Study Region .................................................................................................. 9
1.2 METHODOLOGY................................................................................................. 10
1.2.1 Valley Metro Ridership ................................................................................ 10
1.2.2 Transit Survey................................................................................................ 11
1.2.3 Transit Stop Characterization .................................................................... 12
1.2.4 Geospatial Analysis ..................................................................................... 14
1.2.5 Heat vulnerability Index ............................................................................. 14
1.3 RESULTS AND DISCUSSION ................................................................................. 17
1.3.1 Transit Survey Results ................................................................................... 17
1.3.1.1
1.3.1.2
1.3.1.3

Study Region Walk and Wait Times .............................................................. 18
Valley Metro Bus Stop Characterization Classification Results .................. 21
Valley Metro Ridership and Bus Stop Characteristics ................................. 22

1.3.2 Heat Vulnerability Index ............................................................................. 25
1.4 CONCLUSION ..................................................................................................... 29
2

CHAPTER II: EXPOSURE ........................................................................................... 31
2.1 INTRODUCTION................................................................................................... 31
2.1.1 Research Questions .................................................................................... 34
2.1.2 Experimental Design ................................................................................... 34
2.2 METHODS ............................................................................................................... 37
2.3 RESULTS AND DISCUSSION ........................................................................................ 41
2.3.1 Exposure – Research Question 1 ............................................................... 43
2.3.2 Exposure – Research Question 2 ............................................................... 46
2.3.3 Exposure – Research Question 3 ............................................................... 48
2.4 BROADER IMPACTS ................................................................................................. 50
2.5 CONCLUSION ......................................................................................................... 51

3

CHAPTER III: MITIGATION ....................................................................................... 53
3.1 INTRODUCTION................................................................................................... 53
3.1.1 Assessing Vulnerability to Heat Exposure ................................................. 53
3.1.2 Contextual Vulnerability............................................................................. 54
3.2 VULNERABILITY OUTCOME ....................................................................................... 55
3.2.1 Overview of the Mass Transit System ........................................................ 57
3.3 RESULTS AND DISCUSSION ........................................................................................ 60
3.3.1 Findings Relevant to Vulnerability Outcome........................................... 60
3.3.2 Heat Risk Assessment .................................................................................. 63
3.3.3 General Mitigation Strategies ................................................................... 67
3.3.4 Costs of Recommended Heat Mitigation Strategies ............................. 71
3.3.4.1
3.3.4.2
3.3.4.3
3.3.4.4
3.3.4.5
3.3.4.6

Vegetation....................................................................................................... 73
Cool and Green Pavement........................................................................... 73
Green Walls ..................................................................................................... 75
Green Roofs ..................................................................................................... 76
Solar Panels ...................................................................................................... 76
Water Fountains .............................................................................................. 76
Page 2 of 94

3.3.4.7

Future Innovation ............................................................................................ 76

3.3.5 Fiscal Policy (findings and recommendations) ....................................... 77
3.4 LIMITATIONS ............................................................................................................ 83
3.5 CONCLUSION ......................................................................................................... 84
4

REFERENCES:........................................................................................................... 85

5

APPENDIX: .............................................................................................................. 91
5.1 APPENDIX A: TEMPE TRANSIT SURVEY FOR STUDY REGION ........................................... 91
5.2 APPENDIX B: STANDARDIZED CAP LTER METADATA PACKAGE ASSOCIATED WITH THIS
PROJECT DATA. ................................................................................................................ 93

Page 3 of 94

LIST OF TABLES
Table 1 - Summary of Tempe study region transit stop classification variables. .... 13
Table 2 - Characteristic grading scale for stop vulnerability index. ........................ 15
Table 3 – Final weighting schemes developed by the team members. ................ 17
Table 4 - Walk and wait times from Tempe transit survey. ........................................ 19
Table 5 - Highlights of Tempe Study Region Transit Stop Characteristics and
Amenities. ........................................................................................................................ 22
Table 6 Ten stops with highest vulnerability score across five weighting schemes.
........................................................................................................................................... 26
Table 7– Ten stops with highest vulnerability score per each weighting scheme. 27
Table 8 - Summary Statistics of Thermal Imaging Data ............................................. 44
Table 9 - Average Values of Thermal Imaging Data ................................................. 44
Table 10 - Statistical analysis with outlier removed .................................................... 46
Table 11 – Impacts of Shade type on surface Temperature .................................... 64
Table 12: Transit Survey Likert Scale Matrix .................................................................. 65
Table 13 – Relationship between Human thermal Comfort Index and NOAA Heat
Index ................................................................................................................................. 66
Table 14 – Urban Heat Island Causes and Mitigation Strategies ............................. 68
Table 15 - Costs of Feasible Mitigation Strategies ...................................................... 72
LIST OF FIGURES
Figure 1 - ASU Urban Climate Research Center survey of City of Phoenix Bus riders
indicating perceived Thermal Condition at shaded and unshaded transit stops
during the summer of 2016.............................................................................................. 8
Figure 2 - Transit stop locations in Tempe study region ............................................... 9
Figure 3 - Average daily boardings by season for Tempe transit stops. ................. 10
Figure 4 - Elements that Tempe Transit Rider Respondents Indicated Made Them
Cooler............................................................................................................................... 18
Figure 5 - Minimum Walk Time to Nearest Bus Stop in Tempe Study Region. ......... 20
Figure 6 – Daily Bus Stop Ridership versus Average Daily Estimated Wait Time. .... 21
Figure 7 – Daily Bus Stop Ridership versus Level of Greenery. .................................. 24
Figure 8 – Daily Bus Stop Ridership versus Level of Shading. ..................................... 24
Figure 9 – Stops Identified as Potentially More Vulnerable Due to Lack of Thermal
Comfort Characteristics. ............................................................................................... 25
Figure 10 Map of study region. Green dots represent ValleyMetro stops
contained within the study region. .............................................................................. 35
Figure 11 - Example of thermal image + corresponding normal image. ............... 39
Figure 12 - Average surface temperatures for various configurations of shade ... 44
Figure 13 - Surface temperatures of transit stop components................................. 45
Figure 14 - Best Cooling Methods from Survey Responses ........................................ 48
Figure 15 - Survey responses regarding transit behavior change based on
temperature .................................................................................................................... 49
Figure 16 - Study Area .................................................................................................... 53

Page 4 of 94

PROJECT OVERVIEW
Increasing frequency and severity of heat waves and extreme heat
events in Maricopa County, Arizona, puts residents at risk to heat-related
morbidity and mortality. With expanding urban development and
intensifying urban heat island effect, public transport users are inherently
vulnerable to outdoor heat exposure as they travel and wait for transit.
Research was conducted to investigate how the Tempe Valley Metro
transit infrastructure affects transit use behavior within a selected study
region between Rio Salado Pkwy, S. Rural Rd, E. Apache Blvd, and S. Priest
Dr. Focusing on the study area, a transit survey was conducted to
augment Valley Metro ridership data that lacked transit user thermal
comfort perception and behavioral responses to heat at transit stops. The
eighty-eight responses were linked to various research aspects in this
study. Transit stop amenities were characterized based on the capacity to
provide relief from heat and correlated with transit ridership; the analysis
conveys that stops with higher vulnerability to extreme heat (lack of
greenspace and physical shade) serve a lower volume of passengers.
Geospatial analysis was conducted to assess spatial linkages using
average walk time to the nearest transit stop, average daily ridership,
average wait time, vegetation and shade level for each stop. The results
indicate that most transit stops in the study region are not significantly
vulnerable, though there is concern for stops with moderate ridership and
longer than average wait times. Overall, transit stops on W Rio Salado,
South Priest Dr, South Hardy Dr, and some stops on University Dr are more
vulnerable to extreme heat than average, in terms of lack of vegetation,
direct and indirect shading, ridership, long walk and wait times.

Page 5 of 94

1 CHAPTER I: TRAVEL BEHAVIOR
1.1 INTRODUCTION
Extreme heat events due to climatic changes and excessive urbanization
have been increasing since the beginning of 21st century. In the United
States alone, extreme heat events (EHE) have caused over 3000 deaths
occurring from 1991 to 2003, with arid Arizona leading the nation (CDC,
2006). In Maricopa, there is growing concern that increased frequency
and severity of extreme heat events may put its inhabitants at risk,
especially those needing to be outdoors to travel. Increased extreme
heat events coupled with continued urbanization and the urban heat
island effect worsens heat in metropolitan regions such as Maricopa (Kim,
2007; Tan et al., 2010). A number of studies found that increased heat
waves, increased frequency of extremely hot days, and increased
average temperatures will cause an increase in heat related morbidity
and mortality (Eisenman et al., 2016; Harlan et al., 2013; Saha et al., 2013;
Sarofim et al., 2016; Uejio et al., 2011). From 2003 to 2013 in Arizona, 1,391
deaths were directly caused by heat, while 1,761 were heat related
(AZDHS, 2015). Arizonans age 24 or older have the highest rate of extreme
heat related deaths (Uejio et al., 2011). Heat exposure can cause other
clinical illnesses and can exacerbate pre-existing health conditions such
as respiratory and cardiovascular diseases which increase public health
expenses in a society (Berko et al., 2014; Luber and McGeehin, 2008). In
addition, Karner et al. (2015) argue that outdoor activities and physical
exertion increase the risk of health hazards in an EHE. Therefore, urban
public transport users are inherently vulnerable to adverse heat related
health impacts due to the necessity to walk and wait outdoors.
The Tempe Transportation Master Plan (City of Tempe, 2016) prioritizes
multi-modal transportation and aims to decrease car dependency. Their
goals include improving bus service, expanding the transit network by

Page 6 of 94

introducing additional routes (Orbit Saturn circulator in South Tempe),
developing new modes of transit (Tempe Streetcar Project), and
maximizing biking and walking. However, extreme heat conditions in
Maricopa County may become an impediment to ridership and even
pose health hazards on pedestrians and public transit riders. Moreover,
this problem is exacerbated by the fact that the majority of public transit
users are also socio-economically vulnerable (Fraser and Chester, 2016;
Karner et al., 2015; Valley Metro, 2017). For this reason, it is important to
understand the dependencies between heat, ridership, and the effect of
urban form and amenities at transit stops and streets.
Fan et al. (2016) found that wait times at stops without benches and
shelter are perceived to be at least 1.3 times longer than their actual wait
time, and this does not consider extreme heat. According to Middel et al.
(2016), outdoor thermal comfort is "a complex function of atmospheric
conditions and physical, physiological, psychological, and behavioral
factors", with mostly subjective thermal sensations and responses. Surveys
conducted at City of Phoenix bus stops during the summer of 2016 by the
Arizona State University (ASU) Urban Climate Research Center showed
that 90% of surveyed respondents preferred to be in the shade. At
unshaded bus stops, transit riders reported feeling extremely hot (3.5 times
more often) (Figure 1) and 'very uncomfortable' (twice as much) than at
shaded bus stops.

Page 7 of 94

Figure 1 - ASU Urban Climate Research Center survey of City of Phoenix Bus riders
indicating perceived Thermal Condition at shaded and unshaded transit stops during
the summer of 2016.

The ASU Climate Research Center results showed that physical transit stop
amenities play an important role in rider's comfort and perception of heat.
Valley Metro's 2016 ridership survey did not include any information on
how heat affects rider’s experiences (Valley Metro, 2016). Thus, it is
necessary to explore the technological, behavioral, or physiological
adaptations to heat exposure in Arizona, where EHEs are chronic rather
than episodic (Chuang et al., 2015).

1.1.1 PROBLEM STATEMENT AND RESEARCH QUESTIONS:
Considering the issues with increased morbidity and mortality associated
with high and extreme heat exposure, it is clear that there is an increased
potential for extreme heat to adversely affect transit users in the Tempe
region in the next few decades. This report will look to answer how the
Tempe transit infrastructure affects transit use behavior within the selected
study region. Furthermore, the research investigated whether there are
certain Tempe transit stops or corridors that may be more vulnerable to

Page 8 of 94

heat exposure, as well as any existing correlations between heat
vulnerable locations and ridership.
1.1.2 STUDY REGION
A specific and bounded area of study was selected for this assessment.
Figure 2 below illustrates the 127 Tempe transit stop locations within the
study boundary of Rio Salado Parkway (north border), South Priest Drive
(west), West 13th Street/East Apache Boulevard (south), and South Rural
Road (east). The region includes the ASU Tempe campus, Downtown
Tempe, stadiums, residential housing units and businesses. Thus, the area is
partially highly trafficked (human and vehicular) due to the ASU campus
and inclusion of Downtown Tempe, a popular attraction.

Figure 2 - Transit stop locations in Tempe study region
Bound by Rio Salado Pkwy (north), S. Priest Dr. (west), W. 13th St./E. Apache. Blvd. (south),
S. Rural Rd. (east). Light rail stops are blue triangles and bus stops are green circles (Valley
Metro, 2017).

Page 9 of 94

Based on Valley Metro's Transit System Map (2017), the study transit routes
(124 bus stops inclusive of 3 light rail stops) in Figure 2 would be:


Bus Routes: 30, 48, 56, 62, 65, 66, and 72.



Orbit Neighborhood Circulators with same stops as selected bus
routes: Jupiter, Mars, Mercury, and Venus.



Light rail stops: University/Rural, Veterans Way/Tempe Transit Center,
and Mill/3rd.

1.2 METHODOLOGY
1.2.1 VALLEY METRO RIDERSHIP
To investigate seasonal ridership and transit stop vulnerability, various data
was collected and processed. Valley Metro average daily boardings
were obtained for each stop in the study region across seven seasonal
periods from 2015 to 2016 (Valley Metro, 2017). Figure 3 displays the
Tempe Bus boardings by stop for 2015 to 2016.

Figure 3 - Average daily boardings by season for Tempe transit stops.
Note that this figure only includes Tempe transit stops within the study region. Tempe
stops outside the study region were not included in this analysis.

Page 10 of 94

As of 2015, Tempe has an estimated population of 175,826 (US Census
Bureau, 2017) and Arizona State University's Tempe campus had a fall
2016 enrollment of 51,869 (ASU, 2017). While not all ASU Tempe students
and employees live in the analysis region, a large fraction of them will use
the transit within the study region during the fall and spring seasons. This is
especially true for the light rail, as the light rail route includes many transitoriented developments that have targeted student living. However, bus
ridership fluctuates far less from season-to-season. Although this data is
aggregated at a low temporal resolution, we hypothesize that
temperature has little to no influence on seasonal ridership. It is unclear to
what extent transit rider behavior is influenced by temperature at a dayto-day basis. For example, it is unclear if an extreme heat wave would
cause any noticeable perturbations in ridership. In order to collect a
higher resolution set of data, we conducted a survey and asked Valley
Metro riders about their travel behavior. The transit survey will be
explained in the following section.
1.2.2 TRANSIT SURVEY
Our focus group, with assistance from the rest of the class, put a transit
survey together to investigate the travel behavior and heat exposure of
Valley Metro riders. The survey (available in Appendix A) consists of 10
questions asking about:
1. Demographics of the participants
2. Travel behavior of the participants
3. Heat perception and heat exposure experience of the participants
The students of the UIA class conducted the survey, by going to roughly 30
different bus stops in the study area. We surveyed 88 people during a
short-term heat wave in April 2017 during daylight hours. The average
temperature during the course of survey was 85°F, which is far from

Page 11 of 94

maximum annual temperature in the study area during summer, however,
it was relatively hot for that time of the year. The result of the survey will be
discussed in more detail in section 3.

1.2.3 TRANSIT STOP CHARACTERIZATION
The intent of characterizing transit stops amenities was to identify any
relationships between infrastructure present at transit stops and ridership
to better understand how extreme heat and transit stop infrastructure
impact riders in Tempe. Characteristics were selected based on having
some capacity to impact the degree of experienced or perceived heat
by an individual. A different method for evaluating each characteristic
was used due to the inherent differences between characteristics.
The vegetative greenspace characteristic of transit stops was labeled as
"Green" and rated on a scale from 0-3 with 0 representing no vegetation
or green landscaping infrastructure within 10 feet (3m) and 3 representing
extensive vegetation within 10 feet (3m). This evaluation is subjective
because the stops were rated based upon the presence of greenspace
using transit stop photos from Valley Metro's GeoCenter database (2017).
The rating method for "Green" is the primary limitation of this assessment
due to potential reproducibility issues of results. "Shade" was rated to
identify the presence and type of shading, these options are: no shade,
direct shading from transit structure, indirect tree shade, indirect shade
from building or other gray infrastructure, and both indirect tree shade
and building shade. The purpose of the "Material" characteristic is to
identify paving materials. Understanding the type of paving material can
provide information on heat experienced at transit stops, since different
paving materials have different albedos and heat absorption capacities.
The characterization rating for materials are: concrete, asphalt, and brick.
"Near cool" evaluated the distance (in feet) to the nearest indoor location
Page 12 of 94

that could provide refuge for passengers. This was done using the
measurement feet feature within Google Earth, or alternatively Google
Maps.
The characteristics: "Bench", "Bikerack", "Structure" and "VRT Green" are
ranked as a binary yes or no based on whether they exist at the stop.
Benches were selected because they provide physical comfort and a
place for a person to sit, which can alleviate heat experienced. Bike racks
were included in the study to determine if the presence of bike racks
increases use of a given transit stop. This is partially because the presence
of a bike rack at a stop enhances biking accessibility to a transit stop by
providing a location for riders to leave their bike if they either do not want
to mount their bike on the bus or if mounting is not an option. "Structure"
represents a physical structure built as part of the stop to provide shade
for passengers. "VRT Green" refers to a vertical plant installation that
provides direct shade, such as vines. Both Structure and VRT Green were
considered to investigate the effectiveness of intentional installations for
providing shade. Table 1 summarizes the stop characteristics and their
variable type.
Table 1 - Summary of Tempe study region transit stop classification variables.
Only Tempe transit stops in the study region were classified with these variables. This
inventory of amenities includes 124 bus stops including 3 light rail stops for the given
Tempe study region.

Page 13 of 94

1.2.4 GEOSPATIAL ANALYSIS
Geospatial analysis was performed to determine spatial linkages between
various datasets obtained from Valley Metro, scientific literature and
transit stop classification created by team members from photographic
evidence for the purpose of in-depth analysis of bus stop amenities and
infrastructure characteristics.
Data types and sources used for geospatial analysis:


Valley Metro bus and light rail stops locations and average daily
ridership (2015-2016)



Wait times estimates based on the Valley Metro schedule (Fraser
and Chester, 2016)



Vegetation estimates based on the stop photographs (sourced
from Valley Metro GeoCenter database, 2017) characterization



Shade level based on the stop photographs (sourced from Valley
Metro GeoCenter database, 2017) characterization

Geospatial analysis included walk time to the nearest transit stop,
average daily ridership, average wait time, vegetation and shade level
for each stop.
1.2.5 HEAT VULNERABILITY INDEX
The next phase in the analysis of transit infrastructure and people's
behavior regarding transit in the context of extreme heat was to develop
a heat vulnerability index of all transit stops in the investigated region. This
index combines the data obtained from the transit stop characterization
and the transit survey to determine the degree of heat vulnerability at
each transit stop. The goal of this index is to rank all transit stops based on
their heat vulnerability to determine which stops are the most vulnerable
to extreme heat. The stops were quantitatively rated based on the
characteristics described in Section 2.3. For the vulnerability index, some

Page 14 of 94

characteristics studied in the stop characterization process earlier were
determined to be irrelevant or difficult to account for. For example, the
characteristic "pavement type" was excluded from this study because it
was difficult to evaluate the impact of one particular material, since they
are often used near each other. The characteristic "VRT Green" was
omitted from the index because it was not mentioned in the transit rider
survey, which prevented understanding ridership perspectives on this
feature at transit stops. Table 2 lists each characteristic used and its
associated rating range for stop vulnerability evaluation. As displayed
below, in Table 2, all values range from 0 to 1. The acceptable ranges
were purposely normalized to ensure that all characteristics have the
same impact on overall stop vulnerability rating before incorporation of a
weighting scheme to provide value to each characteristic. The
normalization of results is very important for characteristics such as "Near
Cool", which was scored based on distance to nearest cool location (in
feet), which would have produced complications if the characteristic
remained on a scale of 0-2080 ft, with 2080 feet being the longest
distance to the nearest public cooling relief location. The characteristic
"Greenspace" was slightly modified for the same reason. Tree shade and
structure were separate characteristics that were created out of the
characteristic "shading". The purpose of this was to better represent the
types of questions presented in the survey to better represent transit rider
values in the upcoming weighting scheme.
Table 2 - Characteristic grading scale for stop vulnerability index.

Page 15 of 94

Several weighting schemes were developed to represent different
perspectives and values on heat vulnerability. Each weighting scheme
assigned a level of importance to each characteristic such that
characteristics of greater importance have a larger impact on the
vulnerability score of that stop. The objective of implementing multiple
weighting schemes is to provide stop vulnerability rankings that are
specific to different attitudes and thermal comfort perceptions on heat
vulnerability. In addition, analyzing stop vulnerability form multiple
perspectives will create more defensible results due to the variety of
perspectives being considered. These results have the potential to be
even more compelling if a set of stops are determined to be highly
vulnerable, moderately vulnerable or not vulnerable under all weighting
circumstances.
In order to calculate stop vulnerability rank, the stop characteristic
grading scale and weighting schemes were combined is as follows:


Establish grading scale for each characteristic (between 0 and 1).



Multiply characteristic score by weighting value from given
weighting scheme, summing. all categories for an overall score.



Perform 1-result to determine most vulnerable stops.



The higher the result, the less vulnerable it is.



Repeat for each scheme to obtain ratings based on different
perspectives.

Table 3 displays the five weighting schemes used to rank stop vulnerability.
The scheme "Equal" gave an equal weight to all characteristics, so that
each characteristic had an equal impact on stop vulnerability. The next
scheme, titled "Default 1" was developed by the team members on what
they collectively thought was most valuable for reducing heat
vulnerability, which included high priority to green space and tree shade

Page 16 of 94

(for the urban heat island reduction abilities of plants) and very little
importance on benches and near cool. The scheme "Survey" created a
weighting scheme that reflected the results of transit rider survey
performed by the entire class. The scheme "Bike = NC" was another
scheme developed by the team in which they gave bike racks and
distance to near cool an equally low weight and high value placed on
shading mechanisms. The last scale, titled “Alternative 1” was another
attempt by the team to create an alternative weighting that varied and
could be compared to the previous schemes.
Table 3 – Final weighting schemes developed by the team members.

1.3 RESULTS AND DISCUSSION
1.3.1 TRANSIT SURVEY RESULTS
One question from the Tempe transit survey was identifying transit stop
characteristics that are perceived more favorably for cooling ability by
transit riders. Specifically identifying which characteristics are of higher
priority for keeping cool is important to justify further focuses in the analysis.
Figure 4 shows a summary which elements were reported by transit riders
as keeping them cool. Specifically, respondents favored shade structures
and nearby trees heavily. Respondents were less likely to report benches,
nearby grass, and nearby water as keeping them cool. These findings
provide useful insight of which transit stop characteristics are important to
the perceived thermal comfort and preference of transit riders. In other
words, shade is very important specifically when provided directly by
shade structures or tress, or when nearby trees offer shade. This is an
Page 17 of 94

important component that helps inform which characteristics may be
influential in altering transit behavior.

Figure 4 - Elements that Tempe Transit Rider Respondents Indicated Made Them Cooler.
Note that respondents were allowed to choose as many elements as they felt kept them
cool.

1.3.1.1 STUDY REGION WALK AND WAIT TIMES
To understand riders’ heat exposure and travel behavior, walk times and
wait times were analyzed. First, this was done by examining details of
walking and wait times from the previously discussed Tempe transit survey.
Table 4 shows walk and wait times in a matrix from the survey. Interestingly,
23 respondents had a combined walk plus wait time of 17-30 minutes and
6 respondents had a walk plus wait time of 22 minutes, minimum (Table 4
boxes outlined in red). Although this survey was conducted in over mid
80°F days in March 2017, the questions asked how long these activities
typically took, therefore these behaviors may continue during extreme
heat periods. Specifically, 8 respondents in these 17-minute minimum walk
plus wait time category also indicated that they do not change behavior
at all under extreme heat (while most others said they wait in shade, bring
water, bring umbrellas, etc.). Through these results, it is clear that transit
Page 18 of 94

riders in Tempe – and potentially the greater Maricopa region – could be
vulnerable to extreme heat due to high walk plus wait times to access
transit.
Table 4 - Walk and wait times from Tempe transit survey.
Note that not all transit stops in Tempe or all stops within the study region were sampled.
Stops outlined by the red box are N = 88 total samples.

Walk time analysis for the whole study region was done to spatially identify
regions that are potentially more isolated than others. This geospatial
analysis assumed an average walk time of roughly 3 miles per hour (MPH)
and assumed transit riders walking to a bus stop would follow along the
road network. Figure 5 shows the spatial distribution of walk times to the
nearest bus stop in the Tempe study region. Comparing the walk times to
the nearest stop to the previously reported survey walk times shows that
people report walk times much larger than the greatest times needed to
walk from any point in the study region to a bus stop. In some cases
(about 11 of the 88 respondents), their walk times to a Tempe bus stop
were over double the longest walk time to the nearest stop estimated by
Figure 5 (max of 5 minutes in the figure compared to at least 11 minutes
for some in the survey). This indicates that transit riders are not commonly
walking to the closest bus stop from their origin or perceive their walk to
be longer than it actually is. Regardless of the actual reason why (and the

Page 19 of 94

likely possibility that both are contributing factors to higher reported walk
times), this indicates that transit riders accessing stations are either
vulnerable or not comfortable during hot periods of the day.

Figure 5 - Minimum Walk Time to Nearest Bus Stop in Tempe Study Region.
Note walk times are assumed to be ~3 MPH and follow road networks only (no shortcuts).

To examine wait times further, estimated wait time data from Fraser and
Chester (2016) was combined with ridership data from Valley Metro in a
geospatial analysis to assess if any stops may be more vulnerable due to
high ridership and high wait times. Figure 6 shows the nexus of ridership
and wait times at bus stops in the study region. There are no obviously
high ridership stops with long average wait times. This is not surprising due
to the fact that transit planning aims to serve high volume stops with more
frequent vehicle headways. Although there are not any clearly vulnerable
stops apparent in Figure 6, there are stops that have long wait times that
may be more isolated when considering walk time access (Figure 5)

Page 20 of 94

Specifically stops on W Rio Salado Pkwy, stops on University Dr near Beck
Ave, Farmer Ave, and Maple Ave, and a stop at Priest Dr and 13th St. may
be slightly more vulnerable due to moderate ridership and some longer
than average wait times. However, the bus stops with the longest average
wait times have among the lowest average ridership.

Figure 6 – Daily Bus Stop Ridership versus Average Daily Estimated Wait Time.
Daily ridership correlates to the size of each bus stop node (note the scaling is
non-linear due to the high variance in daily ridership within the region) and wait
times correlate to the color intensity of the nodes (red is high wait time, yellow
around average wait time, and green is low wait time).

1.3.1.2 VALLEY METRO BUS STOP CHARACTERIZATION CLASSIFICATION
RESULTS
A summary of Tempe Transit stops amenities is summarized in Table 5. From
these 124 transit stops, 75% stops are adjacent to some greenery whereas
the remaining 12% provide little to no greenery, which may be concerning
since vegetation is known to alleviate ambient temperatures. The analysis
also found that about 27% transit stops lack significant shading (either
absence of shade structure, canopy shade or indirect shading from
nearby structures/buildings) based on the classification method. Table 5
Page 21 of 94

indicates that concrete is the most common material used at transit stops,
followed by brick and asphalt. Although concrete might be the primary
material, a transit stop may lie in the center of a highly urbanized space
where asphalt may intensify the ambient heat. Most transit stops within the
study region have benches and 56% of stops had nearby bike parking.
Table 5 also conveys the distribution of distance to the nearest indoor
space that may serve as a cooling refuge. The average distance to a
public space is 302 feet (0.06 miles), with a minimum and maximum of 10
feet and 2080 feet (0.4 miles) respectively. Thus, transit stops within the
study region tend to be less than 0.5 miles away from a public cooling
space. However, the study area focuses mostly on Downtown Tempe and
ASU Tempe, which is considered denser than the rest of Tempe in terms of
public facilities. Moreover, the accessibility of some of these public spaces
may change with the season as campus buildings, stores, restaurants etc.
change hours in response to students leaving in summer.
Table 5 - Highlights of Tempe Study Region Transit Stop Characteristics and
Amenities.
Only Tempe transit stops in the study region were classified with these variables.

1.3.1.3 VALLEY METRO RIDERSHIP AND BUS STOP CHARACTERISTICS
To determine if any stops may be classified as vulnerable due to the
combination of high volume of passengers per day and the lack of
thermal comfort characteristics, two more geospatial analysis were done
similar to the analysis in Figure 6. Based on the survey results, nearby trees

Page 22 of 94

and shade structures were the most commonly favored characteristics to
keep riders cool. Additionally, according to the field thermal imaging by
other members in this study, trees and shade structures also provide the
highest level of surface temperature heat relief. As such, the two
characteristics "level of greenery" and "level of shade" at each stop was
compared with daily ridership for each bus stop in the study region to
understand the relationships and potentially identify if any stops my be
more vulnerable. In Figure 7, daily bus ridership is shown versus level of
greenery for each stop in similar fashion to Figure 6. Many of the stops with
the highest ridership also have medium to very high nearby vegetation.
When looking for specific stops that lack greenery with higher ridership,
there are no obvious stops to hone in on (similar to Figure 6). It is clear that
most bus stops near ASU campus have moderate to high greenery
surrounding them. However, there are stops that lack greenery, and they
commonly intersect with stops that also have higher wait times. All stops
on W Rio Salado Pwky have no greenery nearby, and many stops on the
south side of Priest Dr and Hardy Dr also lack greenery. In Figure 8, daily
bus ridership is displayed with level of shade for each bus stop in the study
region. Following similar trends as Figure 6 and Figure 7, the highest
ridership stops are most likely to have direct or at least indirect shade. Yet
again, many low ridership stops also commonly have indirect and no
shade, and many of these stops also were previously identified to
commonly lack greenery and have higher wait times. All stops on W Rio
Salado lack shade, and the same group of stops mentioned above on
the south side of Priest Dr and Hardy Dr also lack shade.

Page 23 of 94

Figure 7 – Daily Bus Stop Ridership versus Level of Greenery.
Daily ridership correlates to the size of each bus stop node (note the scaling is non-linear
due to the high variance in daily ridership within the region) and level of greenery
correlates to the color intensity of the nodes (off-white is no greenery, and dark green is
the highest greenery).

Figure 8 – Daily Bus Stop Ridership versus Level of Shading.
Daily ridership correlates to the size of each bus stop node (note the scaling is non-linear
due to the high variance in daily ridership within the region) and level of shading
correlates to the color of the nodes (no shade is red, indirect shade is gray, and direct
shade is blue). Direct shade is defined shade from a shade structure or trees directly at
the stop.

Although it is apparent that the Tempe study region has high transit
coverage—many stops with adequate shading, greenery, and wait

Page 24 of 94

times—there are many stops that lack all three of these crucial
characteristics. Stops such as those on W Rio Salado, South Priest Dr, South
Hardy Dr, and some stops on University Dr are stops that have higher
vulnerability to extreme heat, however, they also serve a lower volume of
passengers. The specific stops identified as potentially more vulnerable
from the geospatial ridership analysis are highlighted in Figure 9.

Figure 9 – Stops Identified as Potentially More Vulnerable Due to Lack of Thermal Comfort
Characteristics.
No legend is included but note that the same legend from the previous three figures still
applies except the color schemes.

1.3.2 HEAT VULNERABILITY INDEX
Table 3 outlines the rankings of transit stops under the various weighting
schemes developed for this index. The creation of multiple weighting
schemes was an attempt at performing a sensitivity analysis of stop
vulnerabilities to extreme heat. It was determined that some form of
sensitivity analysis was necessary to produce stop vulnerability results that

Page 25 of 94

were defensible. The sensitivity analysis or creation of a heat vulnerability
index was necessary given that none of the team members were experts
in perceived or actual heat vulnerability. By analyzing transit stop
vulnerability from multiple perspectives, it enables the reader to isolate
results that align with their values (for instance, Valley Metro's perspectives
on potential transit stop improvements). More importantly though,
combining the stop rankings from multiple schemes has the potential to
provide even more powerful results in the event that certain stops score
either very poorly, very highly or average across all weighting schemes. As
it turned out, this was the case. There were a handful of stops that scored
extremely vulnerable across all ranking schemes. These stops are Rio
Salado Pkwy & Lakeside Dr, Ash Ave & Rio Salado Pkwy, Forest &
Gammage Pkwy, Rio Salado pkwy at Linear Parl, Mill Ave & Gammage
Pkwy. Ten stops with highest vulnerability score across all weighting
schemes are identified in the Table 6 below. Table 7 shows 10 most
vulnerable stops per each weighting scheme.
Table 6 Ten stops with highest vulnerability score across five weighting schemes.

Page 26 of 94

Table 7– Ten stops with highest vulnerability score per each weighting scheme.

Page 27 of 94

Table 7– Continues

Page 28 of 94

1.4 CONCLUSION
This research investigated infrastructure characteristics of public transit
stops, availability of shading, wait and walk times to determine the
average heat vulnerability of stops and to identify the stops that are at
higher risk than others. Several quantitative and spatial analysis methods
were used to verify the results.
Results showed that majority of the stops have direct or indirect shade
and they are also the stops with higher daily ridership located at
intersections. Unshaded stops are located in-between the intersections
with daily ridership not exceeding 10 people, with the exception of the
stop at Rural Rd & Terrace Rd with up to 22 riders a day. Similarly, stops
with higher ridership and shade availability have medium to high level of
greenery and are also associated with shorter wait times. There is not
enough information to determine the direction of causality between
ridership at stops and the presence of preferred amenities. It seems
intuitive that Valley Metro and the City of Tempe would prioritize
developing (or initially providing) amenities at stops that have (or are
predicted to have) high ridership. Spatial analysis of walk times showed
that majority of the stops are within 5 min reach. However, 13% of
surveyed transit riders reported their walk time to be twice as long.
Considering this difference in walk times, it is certainly plausible that riders
could choose a stop with more amenities, or even plan their trips such
that they are more commonly at better shaded or cooled stops. Future
research should aim to confidently identify the direction of causality
between ridership and stop amenities for regions with extreme
temperatures, which likely also would require higher temporally resolute
data.
It is apparent that stops along W Rio Salado Pwky, Priest Dr and Hardy Dr
are more vulnerable based on the infrastructure and wait time
Page 29 of 94

characteristics. However, these stops have low ridership level which
means that fewer people would get exposed overall. The limitations of this
study did not allow to establish route and cause relationships between
ridership, wait times and available infrastructure, thus, it is recommended
to further investigate this stops and identify the need for improvement.

Page 30 of 94

2 CHAPTER II: EXPOSURE
Extreme heat events are associated with increased morbidity and
mortality. This is especially concerning as all climate models indicate the
Earth's temperature will continue to rise and as the temperature increases,
extreme heat events will become more frequent, more severe, and longer
lasting. This poses an issue specifically in Maricopa County due to its
location in the arid southwest and the additional issues associated with
heavy urbanization such as the urban heat island effect. The heightened
temperatures leave a large portion of the citizens in the county vulnerable
to heat related illnesses and deaths. In particular, during the summer
months, non-automotive transit users in the Phoenix metro area are
especially vulnerable to heat. In combination with a lack of regional
connectivity and disjointed planning of transportation networks, transit
and transportation users are likely to have major parts of their travel
experiences that require non-voluntary exposure to heat. Building from
previous as well as new research, this paper quantifies the thermal
dynamics and characteristics of transit stops and transit stop components
of the Valley Metro bus system. This was accomplished through fieldbased observations of surface and air temperatures of Valley Metro transit
stops in Tempe, AZ. It was found that green infrastructure paired with
shaded structures provided the greatest protection from the heat.
2.1 INTRODUCTION
More than half of the world's population lives in cities, where the
combined effect of urban heat island and summer time extreme heat
events exacerbate already high levels of heat related morbidly and
mortality. Extreme heat is a leading cause of weather related human
mortality in the United States. Heat exposure is a complex problem for
major urban areas. In the arid southwest, extreme atmospheric heat
combined with phenomena like the urban heat island effect can

Page 31 of 94

produce challenges for urban planners and city managers. In particular,
during the summer months non-automotive transit/transportation users in
the Phoenix metro area are especially vulnerable to heat. This is due in
part to the absence of extensive cooling or shade features or corridors in
the Phoenix region. In combination with a lack of regional connectivity
and disjointed planning of transportation networks, transit and
transportation users are likely to have major parts of their travel
experiences that require exposure to heat.
There are several previous studies on extreme heat and human health
that have estimated threshold temperatures where negative health
outcomes become prevalent (Hajat et al., 2006, McGregor et al., 2007),
as well as developed heat preparation plans (EPA 2006), and identify
vulnerable groups within the population (Harlan et al., 2013). While these
approaches offer useful information about the geographic distribution of
potentially vulnerable 'hot spots' in a study region, they are limited in
specifying what components of transit infrastructure mitigate or
exacerbate heat exposure. Karner et al. (2015) illustrates the heat
exposure risk for pedestrians in the Bay area giving some insight into
conditions that urban transit users might be more generally experiencing.
Additionally, links between extreme heat and adverse health outcomes
(morbidity or mortality) are well documented in the literature (Hartmann
et al. 2013, similar studies). An analysis performed by Mills et al. reveals that
mortality from weather related illnesses are projected to increase due to
extreme weather events increasing in frequency and number (Mills et. al.,
2015). Further, there is an important equity dimension to heat vulnerability
and heat exposure– populations (often low socioeconomic status, elderly
age along with isolation, or non-vegetated areas) that are most
vulnerable to heat are also most likely to need to rely on non-automotive
transit for their travel behavior (Harlan et al., 2013). As such, non-

Page 32 of 94

automotive transit experiences are likely the represent the most significant
amount of non-voluntary heat exposure for many urban residents,
particularly those from more vulnerable demographics.
A key step in generating solutions to the problems of heat, health, and
transit-necessitated heat exposure is to understand what the actual
thermal characteristics of transit experiences are. In reviewing previous
research to date, efforts by ASU researchers and ValleyMetro were found
to be largely focused on modeling or simulation approaches (ASU
College of Design). However, these are insufficient as heat experiences
are often a combination of a number of context- or site-specific
environmental and psychological variables not adequately captured in
modeled approaches. Additionally, other research to date in Phoenix has
shown that research about infrastructure or city functioning such as transit
systems in other major US or world cities may not translate well due to the
unique arid land climactic conditions of this region (Weller et al. 2016). As
an example, the Tres Rios Tertiary Wastewater Treatment Wetlands has
been demonstrated by researchers in the Central Arizona Phoenix Long
Term Ecological Research (CAP LTER) program to exhibit unique
hydrologic processes relative to other treatment wetlands in more
traditional, mesic settings (Weller et al., 2015). Importantly, these unique
processes are driven by the incredibly hot, arid climate (Bois et al., 2017).
The need to understand the unique contributions of Phoenix's climate to
person-based heat experiences during transit use is especially pressing
given that climate change will likely exacerbate existing issues related to
heat morbidity and mortality (Transportation Research Board, 2008;
Hondula et al., 2015).

Page 33 of 94

2.1.1 RESEARCH QUESTIONS
As such, the primary goal in conducting this research is to advance the
resolution of the scientific community's understanding of the thermal
characteristics of transit infrastructure. Specifically, this study sought to
quantify surface and air temperatures of transit stops and transit stop
components in the Phoenix metro area. Further, it seeks to understand the
relationship between perceived and actual temperatures at transit stops.
The corresponding research questions are as follows:
1. What are the surface and air temperatures of transit stops and
transit stop components in the Valley Metro bus system?
2. What are the impacts of heat mitigation features (e.g. shade) on
the surface and air temperatures of transit stops and transit stop
components?
3. What are the differences between the perceived and actual
heat relief provided by heat mitigation features at transit stops?
To answer these questions, a field-based approach was proposed in order
to quantify thermal characteristics of transit stops and transit stop
components. The use of instrumentation in the field will allow for a better
approximation of the exact conditions that transit users will experience
during transit-related travel activities.

2.1.2 EXPERIMENTAL DESIGN
Field measurements of different types of surfaces and the air temperatures
of Valley Metro stops were taken for better understanding of personal
heat exposure. The study region is bounded by S Priest Drive and Dorsey
Lane to the west and east respectively, and Rio Salado Dr and Apache

Page 34 of 94

Blvd to the north and south, respectively. A geographic representation of
the study area can be seen in Figure 10.

Figure 10 Map of study region. Green dots represent ValleyMetro stops contained
within the study region.
Red circles denote the 13 stops analyzed in this research.

The study area contained approximately 120 Valley Metro transit stops.
However, within the scope of this project, bus stops were specifically
focused on. Within the study area, 13 bus stops were identified which
represent every possible configuration of transit stop features and
components in the Valley Metro transit system. These include unshaded
bus stops, bus stops with "gray" shade infrastructures (e.g. metal roof), bus
stops with "green" shade infrastructure (e.g. trees), and so on. In addition,
these 13 stops overlap with stops that were identified as vulnerable or
needing particular attention by the Travel Behavior researchers in this
course. While earlier iterations of this research proposed more spatially-

Page 35 of 94

oriented analytical approaches (e.g. transect-based surveys), a sitebased approach was utilized instead. Importantly, this decision was made
in order to prioritize the capture of all possible transit component
configurations at a select number of stops. This will allow the data and
analysis to be broadly applicable and scalable to any bus transit stop in
the Valley Metro system that features stop components observed in this
study. Field research was conducted during Spring 2017. Specifically, all
field observations were made during March and April 2017. Further, field
observations were taken between 1PM and 4PM to capture peak
daytime temperatures.
One of the issues with the given sampling season is that the impacts of
maximum annual temperatures or extreme heat events were unable to
be analyzed. The data and analysis is constrained to this time frame.
However, since data were collected during peak daytime temperatures,
it can be further analyzed for extreme summer temperatures. The 13 stops
were then categorized using a variety of different metrics to classify the
stops components based on the materials, presence and absence and
percent coverage. Further, each of the 13 stops were identified by the
unique 5-digit stop site ID numbers determined by Valley Metro to ensure
consistency in site identification between all project groups of the UIA
class. After identifying the stops, the thermal imaging photos and
temperatures were taken in March and April 2017 during peak hours for
better representation of summer time using infrared imaging (FLIR Exx
Series Infrared Camera) at a consistent distant (3m radius) from each
transit configuration to capture the overall range of these stops as well as
each component’s temperatures.

Page 36 of 94

2.2 METHODS
To address research questions 1 and 2, field-based approaches were
utilized to quantify the various thermal dynamics of transit stops identified
in the experimental design. To address research question 3, the field data
was analyzed in conjunction with data gathered via the course travel
survey (as explained and identified by other project groups) to
understand what heat mitigation techniques are most effective.


Travel Survey Methods

UIA researchers administered travel surveys at different bus stops within
the scope of the project. There were ten questions that were asked. In
order to receive responses free of any bias, each of the survey takers were
able to fill out the form on their own using a pen. This data was then
further organized using Google forms where all of the responses were
consolidated by the UIA researchers. The survey and its associated
questions can be seen in Appendix A and Appendix B.


Field Methods

Field surveys of the 13 transit sub-sites identified previously in the
experimental design were conducted. During field surveys, the goal was
to capture three primary types of data: metadata, thermal data, and
classifications of stop features. The metadata consisted of data that
would help situate and provide context for the field measurements and
thermal images in time and space: these included timestamps, locations,
transit stop ID's, and other relevant information collected at the time of
sampling. The thermal data consisted of thermal characteristics of transit
stops as captured by two field instruments: a FLIR Exx Series Infrared
Camera, and a NCEon SMART Environmental Meter (further known as
environmental meter). Finally, in order to relate thermal data to
configurations of transit stops, the features of each transit stop were

Page 37 of 94

classified. This classification scheme was designed similarly to the scheme
utilized by Group 1 (Travel Behavior) to strengthen synergies between the
datasets. Transit stop components were classified based on material,
presence/absence, and/or percent coverage dependent on the
particular component in question. See Table 8 for details. Importantly, this
schema was extended to all features within a 3m radius of the centermost
point of the stop. Distances were measured in the field using a Keson
165ft/50.2m fiberglass measuring tape. This was done to ensure that the
data and surveys captured the transit stop features that are most likely to
have a direct impact on heat experiences of transit users at the stop.
Table 8 – Classification scheme for transit stop components and materials.

Stop Component

Classification

Pavement

Concrete, brick, asphalt

Seats

Concrete, metal, plastic

Landscaping

Xeric (rocks/dirt), mesic (grass)

Shading

Shade structure, direct tree
shading, indirect tree shading

Protocol for a typical sampling of one of the sub-sites would look as
follows. Upon arrival to the site, researchers power on all equipment and
field instruments. During instrument warm-up, researchers would conduct
the classification of stop features, writing down observations in a field
notebook. The environmental meter was held by a researcher at the
centermost location of the transit stop (defined as the location of the
signpost denoting stop location and information). Ambient air conditions
(air temperature, relative humidity) would then be recorded in the field
notebook as observed through the environmental meter. Finally, a series
of infrared photos were taken using the FLIR camera. The first photo of
Page 38 of 94

each stop was an overview shot, taken while standing approximately 10m
from the stop. This would ensure that all features of the stop would be
captured in one image. Subsequent photos are then taken of each
individual component of the transit stop (pavement, seats, trees, grass,
etc.). These feature-specific photos were taken from much closer
distances that allowed the crosshairs on the FLIR camera to target the
feature directly to capture the surface temperatures. If a feature was
present in both open sunlight some form of shade, then each variation
(shaded/unshaded) was captured independently. Finally, the FLIR
camera produces two images simultaneously upon pulling the trigger – a
normal photo and an infrared thermal image. As each photo was taken,
the image IDs were recorded as metadata in the field notebook to
provide context and ease of identification for both the normal and
corresponding thermal images.

Figure 11 - Example of thermal image + corresponding normal image.
Taken with FLIR Exx Series Infrared Camera.

A proposed component of the field analysis involved the use of iButtons to
capture the impacts of transit infrastructure interactions on personal heat
experiences. This data collected by Spring 2017 UIA course students was
initially proposed to be utilized. These iButtons would have provided air
temperature and relatively humidity measurements recorded at 5 minute
intervals throughout transit interactions. In addition, students also recorded
corresponding information about location, activity, and heat-related
Page 39 of 94

factors that provided context for raw iButton data. However, a significant
challenge encountered was from an informal survey of course members
(and a review of personal data of members of this group) it became
apparent that very few transit interactions were recorded by iButtons or in
corresponding travel diaries. As such, this sample size was deemed too
small to be considered useful. A second challenge encountered was that
microclimate data (i.e. air temperature and relatively humidity) were not
consistently recorded during each site visit. As such, the data for
microclimate at each transit stop were relatively limited compared to
other field measurements.


Data Analysis

Data collected in the field were promptly entered into a spreadsheet for
archival and analysis. Field data and observations recorded in field
notebooks were transcribed into this spreadsheet. Thermal images were
extracted from the FLIR camera via USB, categorized by the subject of the
image (overview, seat, pavement), and stored into corresponding folders.
This choice to organize photos by image subject was done as an
extension of the experimental design – by compiling photos that all
feature the same stop components, a database was created of thermal
images and thermal data for a variety of stop components in a variety of
conditions and configurations.
After data transcription, extraction, and QAQC, data were analyzed using
a variety of tests, statistical analyses, and graphical renderings. To answer
Research Question 1, observations were sorted by stop component, then
by shade type, and then averaged all surface temperature readings for
observations with the same component and shade type. Standard error
was also calculated. This created a dataset of average surface
temperatures for each stop feature as they existed in each of the four
shade classifications utilized. To answer Research Question 2, multiple
Page 40 of 94

linear regression models were run utilizing the R statistical package (CRAN
2017). The regression models were set up so that surface temperature was
tested against each of the factors recorded in the classification of travel
stops, such as stop shade type, percent green cover, etc., as well as
combinations of these factors. These models were designed to tease
apart the significant and/or dominant drivers of the variation in
temperatures measured. Finally, to answer Research Question 3, data was
extracted from the travel survey responses which captured transit users
perceptions related to heat experiences at transit stops. Specifically,
responses to the travel survey questions 6 and 7 were investigated for the
purposes of this study. Question 6 asked, “Is there a temperature at which
you would be uncomfortable waiting at or traveling to a transit stop, and
if so what is it?” Questions 7 asked, “Do any of these elements make you
feel cooler? Select all that apply.” The question then listed out a number
of options such as, nearby trees, nearby grass, shade structure, nearby
water, nearby shrubs, or cool breeze. Respondents’ answers were binned
based on similarities in answers and then number of responses for each
answer type was evaluated as a percent of all total survey responses for
that specific question.

2.3 RESULTS AND DISCUSSION


Thermal Imaging Database

All of the thermal images taken for this project were organized in a
thermal image database created by the Exposure Team. Thermal images
were classified and organized according to the particular feature
(seating, shade structures, etc.) that is the subject of the photo, and
about which the camera captured thermal image. This database is
currently located within the UIA course Box folder, however, these images

Page 41 of 94

will be made into a publicly accessible repository in the future. This
database of thermal images has been structured to allow future work,
both within and beyond the scope of this particular UIA course section, to
expand on these initial results.


Climate & Heat Dataset

A dataset of temperatures extracted from raw thermal imaging files,
along with corresponding time, location, and climate data were
collected for each of the bus stops explored. The metadata identifies
when the photo was taken, the researchers who were involved in taking
the photo, the transit mode that the stop represents, the unique image ID
and thermal image id, and what the subject of the photo is. Each photo
in the thermal imaging database has been given an individual sample ID
number. All of the sample ID's are easily identified by other
observers/researchers for their respective stops and attributes due to that
each sample ID has a corresponding Stop ID and NextRide ID that have
been cross-referenced from Valley Metro's database of stops.
Within the thermal image, there was a timestamp, as well as readings of
image subject surface temperature (°F), and a range of minimum and
maximum temperatures within the photo (°F). At the time of the images
being taken, the existence of shade as well as the type of shade was
recorded. Each thermal image taken in the field was organized in
categories based on the transit feature that is subject of each photo.
These include the following possible subjects: overview shots,
pavement/flooring, grass, stop structures (walls, poles, etc.), and heat
mitigation features (built or green shade components). This was done to
ensure that current and future research would be able to analyze the
comparative surface temperatures of various transit components
(overview photos), in addition to the variability in surface temperature

Page 42 of 94

profiles of individual stop features (various other subject photos). In
addition to photo-specific data, an catalog was created of heat-related
stop features at each sampling site (e.g. transit stop) where images were
taken. Importantly, the classification scheme for this catalog was
designed to align with the stop classification scheme utilized by UIA Spring
2017 Group 1, enabling potential synergies between the biophysical data
and the behavioral data. See Appendix B for a complete metadata
package.
2.3.1 EXPOSURE – RESEARCH QUESTION 1
Figure 11 above is an example of a thermal image taken in the field, and
represents the primary method utilized for measuring surface
temperatures at transit stops. A key feature of this image is that it
captured a range in surface temperatures of 48°F. This large range in
surface temperatures was fairly representative of what was found
throughout the study: surface temperatures of transit components varied
greatly. This variation was seen both across various components (e.g.
pavement vs chairs). As an example, the shaded areas of pavement at a
transit stop came in at around 85°F, while pavement areas just outside of
that shade showed a extreme increase to 110°F. This quickly guided the
analysis to be broken down by shaded vs. unshaded components, as well
as looking at each component separately. Tables 8 and 9 display some
summary statistics about the thermal data collected. Following the tables,
Figures 12 and 13 show a visual representation of the data collected. From
these results, it is clear that shade has a noticeable affect on the surface
temperatures, but more importantly, it is clear that tree shade has the
greatest effect.

Page 43 of 94

Table 8 - Summary Statistics of Thermal Imaging Data

Table 9 - Average Values of Thermal Imaging Data

Figure 12 - Average surface temperatures for various configurations of shade

Page 44 of 94

Figure 13 - Surface temperatures of transit stop components.
Note large differences between minimum and maximum temperatures both within and
between different transit components.

Figure 12 breaks down transit stops by components and shows the
average, maximum, and minimum temperatures of those separate
components. The first point to be made of the data collected is that the
components had a range from 69.1°F to 124°F.This 54.9°F difference made
it evident that the components used at and around the stop are
extremely important when it comes to heat mitigation. Another interesting
point drawn from this table, is that grassy areas had the lowest average
temperature at 79.5°F. This temperature was 19.1°F cooler than the next
lowest component, making this by far the coolest component at any of
the transit stops. One outlier of data, 90.3°F when shaded, could be
removed from this analysis of grass. This outlier was high because this
specific grassy area was more patchy than most and had only about 40%
grass coverage. As such, this particular thermal image sample was found
to capture a surface temperature that was not representative of what
actual surface temperatures likely were. The results thus changed Grass &
Shaded Subjects to 76.8°F average, 85.7°F max for Grass & 89.39°F
Page 45 of 94

average for Shaded Subjects. This new data is shown in Table 10. This
would increase the gap from grass to the next coolest component to a
change of 21.8°F. As it pertains to shade the average temperature
changed from a manageable 89.4°F average for shaded components to
a extreme 112.0°F average for unshaded ones, yielding a temperature
change of 22.6°F, which brought shade into first place as the leading
transit stop cooling characteristic. This did not distinguish between green
or grey shade though.
Table 10 - Statistical analysis with outlier removed

2.3.2 EXPOSURE – RESEARCH QUESTION 2
According to the results, the stops with the largest difference in
temperature were stop 860 (50°F difference) and stop 859 (61°F
difference). Taking this into consideration, it is clear the temperature
differences varied by feature/surface material. As the research indicates,
the stops had variable amounts of shade to consider in regards to
temperature (no shading, direct shading from transit, structure (awning,
Page 46 of 94

etc.), indirect tree shade, direct tree shade). Areas that were concrete
and shaded showed little difference in gap temperatures compared to
those areas that were concrete and unshaded. For example, the shaded
concrete area at bus stop 825 had a range of 37°F while the unshaded
concrete area at stop 860 had a range of 38°F. However, seats that were
shaded had a more substantial range than those that were unshaded. For
example, the shaded seats at stop 859 had a range of 44°F while the
unshaded seats at stop 3376 had a difference of only 41°F.Another factor
that considered in the data was the grass at each stop. The grass in the
shaded area at stop 859 had a range of 54°F while the grass in the
unshaded area at the same stop had a range of only 47°F. Overall, the
areas that were under shaded conditions showed a larger gap in the
temperatures than those that were unshaded.
To more broadly analyze the relationship between transit stop features
and surface temperatures, multiple regression analysis as utilized.
Measured surface temperature data were regressed against transit stop
classifications (e.g. presence/absence of shade, percent green cover). Of
all variables tested, only shade type was significantly related to surface
temperatures (p < .001). Shade type included the categories of: no
shade, built shade, direct or indirect tree shade. Further, shade type
explained the majority of the variance in our surface temperature data (r2
= 0.69). There were not any significant correlations or relationships with air
temperature data. This is likely due to the smaller than anticipated sample
size, as many of the field data did not feature complete or adequate
microclimate data.

Page 47 of 94

2.3.3 EXPOSURE – RESEARCH QUESTION 3
A portion of the field data was compared to the heat perception
gathered by the transit surveys. One question proposed by the survey
was, "What transit stop elements make you feel cooler?" The responses are
displayed in Figure 14 below. 84% of the survey respondents perceived
that nearby trees provided significant heat relief. And 74% of respondents
replied that shade structures provide heat relief and a large portion of
respondents said both provided heat relief. The field data measurements
were explored to see if this perception was true. It indicated that tree
shading provided on average greater heat mitigation versus built shade
structures. Grey shade structures averaged about 100°F, tree shaded
areas averaged about 95°F, and a combination of the two brought the
temperature down to about 90°F. This reinforced the perception of the
surveyed public, potentially guiding any mitigation strategies. These
comparisons are also important as heat stress and heat-related health
outcomes are a combination of context- and site-specific environmental
and psychological conditions/stressors.

Figure 14 - Best Cooling Methods from Survey Responses

Page 48 of 94

The results of Figure 14 and the thermal data collected are important to
consider, especially as it relates to the survey results for the question, "Is
there a temperature at which you would be uncomfortable traveling to or
waiting at a transit stop?" This question sought to see at what temperature
transit users behavior changed. Based on the results shown in Figure 15, a
large portion of transit users begin feeling uncomfortable at about 90
degrees and a portion of the transit users do not have a choice and must
continue using transit no matter what the temperature. While the
temperature data presented here is limited to surface temperatures,
these can be used as a proxy for general heat conditions at transit stops.
It is well known that for a large portion of the year, the Arizona
temperatures exceed these temperatures respondents chose.
Additionally, many of the surface temperatures measured here during
Spring are already exceeding these thresholds. This means that a large
majority of transit users are uncomfortable or stop using transit when the
temperatures thresholds are exceeded, therefore, it is important to find
the best ways to help mitigate heat at these stops in order to help make
the rider experience not only more bearable, but safer.

Figure 15 - Survey responses regarding transit behavior change based on
temperature

Page 49 of 94

2.4 BROADER IMPACTS


Scientific

Despite the desire to provide a wide-ranging understanding of how
infrastructural thermal characteristics affect human travel, the scope of
this study is too limited. As such, the data and results were formatted in
line with data management best practices to facilitate future research
into this topic. Specifically, the data and metadata are formatted for
compatibility with the Central Arizona Phoenix Long Term Ecological
Research (CAP LTER) database. The hope is that further research can
continue to expand upon the thermal image and heat databases to
provide a more comprehensive overview of the thermal characteristics of
transportation infrastructure. In particular, this research would be most
relevant to decision makers and practitioners if data from summer seasons
were able to be collected and analyzed to better highlight how the
findings here relate to the performance of these infrastructure
components during extreme heat events. With this in mind, extensive
metadata has been included in the report to ensure the dataset is easily
accessible, reproducible, and expandable.


Societal

Considering the drastically changing climate and the corresponding
effect it will have on temperatures in the Phoenix-metro area, this study
sets a powerful baseline for residents and municipal leaders looking to
improve transit-related infrastructure. With the recent trend of locally
funding transportation projects through voter initiatives, citizens across the
nation are becoming a more integral part in the transportation planning
process (Goldman et al, 2001). Furthermore, many municipalities
implementing these locally funded projects are required to present a
complete list of features and plans before the funding initiative can go for

Page 50 of 94

a vote. As such, local governments and transit agencies are becoming
more cognizant of providing features that appeal directly to the users, lest
the initiative fail. By empowering the community with the information
contained in this report, residents will be able to demand certain materials
or features be present in the transportation plan before it becomes fully
implemented.

2.5 CONCLUSION
Research to date has demonstrated the growing concerns and impacts
related to heat exposure in urban areas. In particular, negative health
outcomes (i.e. morbidity, mortality) related to heat have been shown to
be the predominant weather-related health stressor in Maricopa County.
One of the most common non-voluntary sources of heat exposure for
people in urban areas is transit use. Not coincidentally, people that are
dependent on transit often are also demographic categories that recent
research has identified as vulnerable. The thermal dynamics,
characteristics, and processes of transit experiences that create health
outcomes are not well understood or quantified. To create a better
resolution of this information, a field-based approach was utilized to
quantify surface temperatures of transit stops and transit stop components
in the Phoenix metro area. Field observations at 13 bus stops in the City of
Tempe yielded data on the surface temperatures ranges of entire transit
stop assemblies, as well as more detailed information on the thermal
characteristics and dynamics of specific transit stop components.
Transit stop features were classified by presence/absence, type,
and material. Statistical and regression analyses revealed that, among all
variables measured, shade type provided that greatest and most
significant explanation for variance observed in surface temperature
data. Specifically, tree shade (both direct and indirect) were found to
Page 51 of 94

have a stronger cooling effect than built shade structures. Generally
though, a combination of both built and tree shade resulted in the most
dramatic cooling effects, with temperatures ranges at some stops as
great as 55°F. Although this study was conducted during the relatively mild
climate conditions of Spring, they provide valuable insight into the
broader thermal dynamics and interactions at play at transit stops.
Importantly, these results and data are generally applicable to nearly all
transit stops in the ValleyMetro network, as stops surveyed in this study
contained all possible combinations of bus stop components.

Page 52 of 94

3 CHAPTER III: MITIGATION
3.1 INTRODUCTION
Group 1 (Transport) and Group 2 (Exposure) identified and assessed the
effects of heat exposure on the behavior of users of mass transit in urban
environments by focusing on a geographical area in Tempe, Arizona
bounded on the north by Rio Salado Parkway; on the east by Rural Road;
on the South by 13th Street/Apache Boulevard; and on the West by Priest
Drive. Figure 16 show, a map of the study area follows:

Figure 16 - Study Area

Group 3 (Mitigation) reviewed the findings of Group 1 (Transport) and
Group 2 (Exposure) and prepared this chapter that describes the findings
and proposed recommendations to reduce the impact of heat exposure
on mass transit riders.
3.1.1 ASSESSING VULNERABILITY TO HEAT EXPOSURE
O’Brien et al. (2007) distinguish between “outcome vulnerability” and
“contextual vulnerability” in the discourse surrounding climate change.
These two interpretations frame vulnerability in very different ways.
Outcome vulnerability frames issues around science; contextual
vulnerability assesses climate change from a human and security point of

Page 53 of 94

view. Outcome vulnerability focuses on future climate change;
contextual vulnerability privileges current climate variability. Outcome
vulnerability suggests quantitative methods like dose response models
while contextual interpretation looks to qualitative methods like case
studies and surveys.
The tool provided by O’Brien et al. (2007) includes possible responses to
the following dimensions located on a spectrum: prioritized questions,
focal points, methods, identified results, and policy responses. Applying
the tool to Eisenman et al. (2016) and Jenerette, et al. (2011) provides a
useful way to frame vulnerability to heat exposure.

3.1.2 CONTEXTUAL VULNERABILITY
Eisenman et al. (2016) uses weather data from 2005-2010 in Maricopa
County to analyze the mortality of four groups that exhibit tendencies that
have been shown “to contribute to social vulnerability during heat waves”
(91). Eisenman et al. (2016) considered “access to public cooled spaces,
home air-conditioning and thermal protection of residential buildings” as
“infrastructural variables.” Eisenman et al. (2016) defines heat events
“summertime temperatures that are substantially hotter and/or more
humid than the historic average for that location at that time of year” (p.
92).
The study produced a “data set of 49,394 census-tract/days … [and]
yielded 1584 deaths from all-internal causes (814 on heat days and 770 on
control days) and 40 deaths from heat-related illness (35 on heat days
and 5 on control days)” (93). Eisenman et al. (2016) extracted four factors
that “explained 77% of the variability in the dataset” (p. 93).
1. Socioeconomic vulnerability based no health insurance, working
outdoors in construction, and female headed households.
Page 54 of 94

2. Renters living alone without a vehicle.
3. Older aged female living alone.
4. Agriculture or extraction industries (p. 93)
The approach of Eisenman et al. (2016) is consistent with contextual
vulnerability as defined by O’Brien et al. (2007) because it focuses on
human and security aspects related to the mortality of four socially
vulnerable groups caused by the interaction of extreme heat and
infrastructure. In addition, Eisenman (2016) address the following issues in
ways that are consistent with O’Brien et al. (2007):
1. Prioritizes questions related to a region (Maricopa County) and
special social groups (the socially vulnerable);
2. Focuses on current climate variability based on historic weather
data in Maricopa County from 2005-2010, not future projections;
3. Identifies institutional and socio-economic constraints on local
responses by modeling the availability of the three infrastructural
variables to the socially vulnerable;
4. Describes policy responses including local constraints (e.g. access
to cooled spaces); and
5. Concludes that the identified “adaptive resources to extreme heat”
had “mixed results” (p. 98).

3.2 VULNERABILITY OUTCOME
Jenerette, et al. (2011) offers a perspective more consistent with outcome
vulnerability as described by O’Brien et al. (2007). Consistent with the
prioritized question of the outcome vulnerability, Jenerette, et al. (2011)
analyzed “three decades (1970–2000) of land surface characteristics and
residential segregation by income in the Phoenix” to assess whether
Page 55 of 94

human modification of urban vegetation affected ecosystem cooling
services (2637) and, in turn, exacerbates climate change. Jenerette, et al.
(2011) asserts that vegetation may cool surfaces by 25 degrees C and the
loss of habitat now may contribute to higher temperatures in the future.
Jenerette, et al. (2011) develop a forward-looking approach called the
“urban heat riskscape,” which is an integrated assessment model that
shows “spatial variation in risk exposure and potential human vulnerability
to extreme heat” (2637). “Urban heat riskscape” measures gains and
losses associated with identified results of outcome vulnerability. Further,
the article asserts that in order to “to alleviate neighborhood inequality in
risks from extreme heat through increased vegetation and evaporative
cooling, large increases in regional water use would be required” (2637);
this provides a metric to assess whether the increased demand on scarce
regional water resources would constitute an inappropriate action under
the outcome interpretation.
By advocating “a systems evaluation of the benefits, costs, spatial
structure, and temporal trajectory for the use of ecosystem services to
moderate climate extremes” (2637) and suggesting that “increasing
vegetation is one strategy for moderating regional climate changes in
urban areas and simultaneously providing multiple ecosystem services”
(2637), Jenerette, et al. (2011) assume an evenhanded, forward-looking
orientation associated with outcome vulnerability.
Because this project focuses on the effects of heat on the average user of
transit stops, it is more consistent with the outcome vulnerability described
above. In this vein, Group 3 (Mitigation) evaluated strategies to mitigate
the heat exposure of transit stops and recommends that Valley Metro
plant more vegetation, especially trees, near transit stops. However,
Group 3 (Mitigation) believes that the project should be extended to

Page 56 of 94

assess issues having to do with contextual vulnerability in ways calculated
to reduce the impact of heat on vulnerable populations like the elderly,
children, or homeless who are most likely to use the transit system during
extreme heat events and suffer poor or tragic outcomes.

3.2.1 OVERVIEW OF THE MASS TRANSIT SYSTEM
In 2004, voters approved Proposition 400, which authorized continuing for
another 20 years a half-cent sales tax to fund improvements to freeways,
streets, and the mass transit system in Maricopa County as part of the
Regional Transportation Plan (RTP) (Sjoberg Evashenk Consultants, 2016, p.
1).
Pursuant to Proposition 400 and state law, Sjoberg Evashenk Consultants
submitted an audit in November 2016 entitled, Performance Audit:
Maricopa Association of Governments Regional Transportation Plan. The
information contained in the audit is reliable due to the high standards
that auditors must meet. Although the project only addresses a small,
bounded area in northern Tempe, it is important to understand the broad
context in which the transit system operates.
Partners in the RTP are:


Maricopa Association of Governments (MAG);



Arizona Department of Transportation (ADOT);



Valley Metro Regional Public Transportation Authority (RPTA);



Sixteen local jurisdictions including Phoenix, Scottsdale, and Tempe,
11 towns, and three Native American Communities;



Maricopa County; and



Valley Metro Rail, Inc. (Sjoberg Evashenk Consultants, 2016, p. 8)

Page 57 of 94

RPTA “is a political subdivision of the State of Arizona” and is “responsible
for short-term planning and the operation of the regional bus transit
system in Maricopa County” (Sjoberg Evashenk Consultants, 2016, p. 8).
Valley Metro Rail, Inc. is “a non-profit, public corporation formed by the
cities of Glendale, Mesa, Phoenix, Tempe, and Chandler” and is
“responsible for the design, construction, and operation of the light rail
system in Maricopa County” (Sjoberg Evashenk Consultants, 2016, p. 8).
Valley Metro Rail, Inc. was “administratively combined” with RPTA under
single Chief Executive Officer. Together Valley Metro Rail, Inc. and RPTA
are known as Valley Metro (Sjoberg Evashenk Consultants, p. 8). In the
rest of this report, Group 3 (Mitigation) will use the term “Valley Metro” to
identify the entity that manages the light rail and bus systems. Most of
transit stops within the study area are bus stops although a short segment
of the light rail system is included on the northern boundary of the study
area.
Valley Metro manages the following relevant transit services:


Fixed Route: Bus service provided on a repetitive, fixed schedule
basis along a specific route.



Bus Rapid Transit (BRT): Fixed route bus mode where a majority of
each line operates in a separated right-of-way dedicated for public
transportation use during peak periods and includes features that
emulate services provided by rail, such as defined stations and
traffic signal priority.



Express / RAPID: Bus service intended to run faster than normal bus
services between specific commuter or destination points.



Supergrid: Bus service providing frequent local service throughout
the day and evening.

Page 58 of 94



Light rail/high capacity: Urban rail system that uses light rail vehicles
on fixed rails in a dedicated right-of-way space.



Paratransit (dial-a-ride): Transit service that is more flexible than
conventional fixed route service, including dial-a-ride services
throughout Maricopa County. These services provide Americans
with Disabilities Act paratransit service to individuals with disabilities
that cannot access the fixed route bus and rail system (Sjoberg
Evashenk Consultants, 2016, p. 60).

Valley Metro light rail outperformed the six-peer average for four of the
five categories including:


Operating Cost per Revenue Mile;



Boardings per Revenue Mile;



Farebox Recovery;



Operating Cost per Boarding; and



Subsidy per Boarding.

In 2014, the subsidy per boarding was 84 percent below six peers (Sjoberg
Evashenk Consultants, 2016, p. 62). The six peers were: Dallas, Denver,
Houston, Sacramento, San Diego, and Utah (Sjoberg Evashenk
Consultants, 2016, p. 90).
As regards light rail, Sjoberg Evashenk Consultants (2016) found:
With the exception of on-time performance, we found favorable trends
across each of the performance metrics. For instance, total boardings
increased by nearly 12 percent from fiscal year 2010-2011 to fiscal year
2014-2015, from nearly 12.8 million boardings to 14.3 million
boardings. While total boardings increased over the audit period,
boardings have remained fairly constant over the last three years. Other

Page 59 of 94

performance measures indicate the strength and operation of the system,
such as the farebox recovery that shows the portion of total operating
expenses covered by passenger fares. In Maricopa County, the light rail
system’s farebox recovery increased by more than 24 percent over the
five-year audit period from nearly 33 percent in fiscal year 2010-2011 to 41
percent in fiscal year 2014-2015 (p. 63).
As regards fixed route bus transit, Sjoberg Evashenk Consultants (2016)
found:
Between fiscal years 2010-2011 and 2014-2015, fixed route bus system
performance showed improvements in certain areas and slight decreases
in performance in other areas …. For instance, we noted improvements in
five of the 12 metrics reviewed – total boardings, boardings per revenue
mile, safety incidents, security incidents, and mechanical failures. While
fixed route bus service outperformed peers as discussed earlier in this
chapter, metrics related to service efficiency such as operating cost per
revenue mile and operating cost per boarding showed an increase
between fiscal years 2010-2011 and 2014-2015. With costs increasing and
the average fare remaining relatively consistent, the farebox recovery
declined by nearly 7 percent and the subsidy per boarding increased by
more than 10 percent (p. 65).
3.3 RESULTS AND DISCUSSION
3.3.1 FINDINGS RELEVANT TO VULNERABILITY OUTCOME
Sjoberg Evashenk Consultants (2016) made the following findings that are
relevant to mitigation of heat exposure and associated outcome
vulnerability:


Congestion trends on freeways and arterial streets has generally
increased between calendar years 2011 and 2014, although
average speeds have slightly increased on freeways and slowed on

Page 60 of 94

arterial streets – nonetheless, these trends are better than those
reported nationally (p. 2).


Light rail and bus transit continue to outperform peer agencies in
terms of boardings, farebox recovery, and operating costs; transit
vehicles were on-time at least 92 percent of the time or higher
between fiscal years 2010-2011 and 2013-2014 and mechanical
failures have been decreasing over the same timeframe [citations
omitted] (p. 3).



Some active traffic management tools, such as traffic signal
synchronization and transit signal priority, are in place to enhance
mobility. However, the region is still moving toward more active
real-time monitoring and dynamic, proactive adjustment of the
system to meet traffic conditions on a 24-hour a day basis (p. 3).

Systemwide ridership was five percent higher in fiscal year 2014-2015 than
in fiscal year 2010-2011; however, it “declined from a high of 73.4 million
boardings in fiscal year 2012-2013 to nearly 70.8 million boardings in fiscal
year 2014-2015, a decrease of more than 2.6 million boardings or nearly 4
percent” (Sjoberg Evashenk Consultants, 2016, p. 60). The decline was
attributed to:


A better economy and lower prices for gas, and



A declining number of students from Arizona State University. The
number of annual passes decreased by 55 percent from 18,000 in
2009-2010 to 8,000 in 2013-2014, which translates into 15,000 student
trips per day. Valley Metro attributes the decline to online courses,
free shuttles, increased prices for passes, and more student housing
near campus. (Sjoberg Evashenk Consultants, 2016, p. 60).

A survey of riders in 2014 showed that although 75 percent of riders
expected to continue usinging the transit system, the satisfaction rate
Page 61 of 94

slipped slightly to 70 percent. Reasons cited by riders for decreased
satisfaction included the “need for increased service frequency or longer
hours of operation, lack of on-time performance, and rude or
unprofessional drivers” (Sjoberg Evashenk Consultants, 2016, p. 60). Riders
did not cite exposure to heat or lack of features intended to reduce the
exposure to excessive heat as reasons for dissatisfaction.
Funds from Proposition 400 are “dedicated to projects approved and
proposed as part of the original 2003 RTP” including:


344 total miles of new or improved freeways and highways;



275 miles of new or improved streets;



34 major intersections;



27.7 new miles of light rail; and



40 enhanced or new bus routes (Sjoberg Evashenk Consultants,
2016, p. 5-6).

Funds collected pursuant to Proposition 400 must be spent as follows:


Freeway: 56.2 percent for freeways including capital expense and
maintenance.



Arterial: 10.5 percent for major arterial street and intersection
improvements.



Transit: 33.3 percent for capital construction, maintenance, and
operation of public transportation as well as for capital construction
costs associated with a light rail system (Sjoberg Evashenk
Consultants, 2016, p. 6).

About two-thirds of the funds generated by Proposition 400 are
earmarked for freeways and streets with the remaining third devoted to
transit. The open question is whether this is an appropriate allocation.

Page 62 of 94

Efforts to reduce the effect of extreme heat on transit riders could be
enhanced if the political will existed to change the proportion spent on
transit.
When Proposition 400 passed in 2004, projections anticipated that $14.3
billion would be generated over twenty years from 2006 to 2026; however,
the economic recession from 2007 to 2009 reduced projected sales tax
revenues to $8.6 billion, or 40 less than the amount originally projected. A
shortfall of this magnitude constrains the options and resources that Valley
Metro can devote to efforts intended to reduce the effects of extreme
heat on transit riders.
Group 3 (Mitigation) finds it intriguing that in an audit comprising 129
pages, no mention of extreme heat or climate change or global warming
could be found by electronic searches. Based on this finding and the
conversation with a representative of Valley Metro, heat and its effects on
riders and infrastructure may not be a high priority. Group 3 (Mitigation)
believes that outcome vulnerability of Valley Metro has increased due to
decreased ridership, less satisfied riders, misallocation of tax revenue that
favors construction of roads over transit, and declining revenue that has
not met expectations that arose around the passage of Proposition 400.

3.3.2 HEAT RISK ASSESSMENT
The purpose of the Heat Risk Assessment is to determine which Valley
Metro bus stops in the study area are in the greatest need of heat
mitigation intervention. Heat mitigation strategies include shading, cool or
green pavement, and cool or green roofs. To recommend where to install
heat mitigation strategies, Group 3 (Mitigation) devised a tool that uses
data from the Transit Survey and measurements from thermal images.

Page 63 of 94

Member of Group 1 (Transport), Group 2 (Exposure), and Group 3
(Mitigation) administered the Transit Survey throughout the study area. The
survey asked questions regarding the travel behavior of riders including
perceptions of heat exposure. Two principal findings from the survey are
addressed in the Heat Risk Assessment: average walk time and average
wait time. On the survey, respondents were asked, “how long do you
typically wait at a transit stop?” and given four options: ‘1-5 minutes,’ ‘6-10
minutes,’ ‘11-20 minutes,’ or ‘over 20 minutes.’ Respondents were also
asked, “how long did it take you to reach this transit stop?” and were
given the same four options, plus a ‘I transferred to another line’ option.
Thermal images were collected by Group 2 (Exposure) at thirteen transit
stops in the study area. The images recorded the surface temperatures of
the following infrastructure elements at the stops: shaded concrete, unshaded concrete, shaded seats, un-shaded seats, shaded metal seats,
un-shaded metal seats, shaded grass, un-shaded grass, un-shaded
asphalt, and un-shaded brick. Some of the results are embodied in table
11, which shows the average, maximum, and minimum surface
temperature readings of the infrastructure element at the thirteen stops.
Table 11 – Impacts of Shade type on surface Temperature

Page 64 of 94

The Heat Risk Assessment combines the Transit Survey Results with data
from the thermal images; however, the measurements are not expressed
in the same units. The Transit Survey measures time and the thermal
images measure temperature. To address this inconsistency, the
measurements were converted to four-level Likert scales. Because the
average walk and wait times were already divided into four response
categories, they were converted to a Likert scale with values from one
through four categories on a spectrum from ‘1-5 minutes’ to ‘over 20
minutes.’ The Transit Survey Likert scale matrix is presented in table 12.
Table 12: Transit Survey Likert Scale Matrix

As a test case, Group 3 (Mitigation) chose to analyze the transit stop at
the intersection of University Drive and College Avenue (UDCA).
According to the Transit Survey data, the average walk time the UDCA
stop was ‘11-20 minutes,’ and the average wait time was ‘11-20 minutes.’
Therefore, by using the Transit Survey Likert Scale Matrix, the UDCA stop
scored a six. To keep the values on a four-level Likert scale, six was
divided by the number of values (n = 2) to calculate an overall score of 3
out of 4.
To convert the thermal image data to a four-level Likert scale, the
average measured temperatures of each of the infrastructure elements
was applied to a Heat Index prepared by the National Oceanic and
Atmospheric Administration (NOAA). The NOAA Heat Index presents four
different temperature ranges associated with different levels of

Page 65 of 94

caution/danger and heat disorders. To convert these temperature ranges
to a four-level Likert scale, the values one through four were assigned to
each temperature range respectively, with one being equal to the 26.7 31.7℃range and four being equal to the 54.4℃or higher, please refer to
table 13. The NOAA Heat Index Likert Scale Matrix is presented below. An
extra category with a Likert value equal to zero was also added for
temperatures of less than 26℃
.
Table 13 – Relationship between Human thermal Comfort Index and NOAA Heat Index

Page 66 of 94

At the UDCA stop, the Group 2 (Exposure) found shaded concrete,
unshaded concrete, and an unshaded seat present. Using the NOAA
Heat Index Likert Scale Matrix, these infrastructure elements have a
combined Likert scale value of seven. To keep the values on a four-level
Likert scale, seven was divided by the number of values (n = 3) to
calculate an overall score of 2.33 out of 4. To find the Heat Risk
Assessment for the UDCA stop 3 out of 4 is added to 2.33 out of 4 with the
result totaling 5.33 out of 8.
Group 3 (Mitigation) recommends that Valley Metro expand the Transit
Survey to its entire service area and take thermal images of all stops so
that Valley Metro can calculate the Heat Risk Assessment for all transit
stops that it operates. Using the Heat Risk Assessment tool to calculate
the scores for all transit stops will allow Valley Metro to identify to most “at
risk” stops. Depending on the availability of funds, Valley Metro should
create a plan to retrofit the most “at risk” transit stops with appropriate
mitigation strategies. These steps would allow Valley Metro to reduce the
outcome vulnerability of its system and provide a better experience for its
riders.

3.3.3 GENERAL MITIGATION STRATEGIES
Although changing urban form will provide some relief to transit riders over
the long term as they walk to and from stops, most of the strategies that
reduce or eliminate urban heat islands will not bring immediate relief to
transit riders sweltering at increasingly hot transit stops in
Phoenix. Yamamoto (2006), for instance, includes a comprehensive list of
mitigation strategies relevant to urban heat island (UHI) at all levels and
scales, but most are not relevant to transit stops. Mitigation strategies
intended to reduce the causes and effects of urban heat islands involve

Page 67 of 94

significant modification to the urban form; however, the responses are
implemented at scales larger than transit stops and over extended
periods of time – years, if not, decades – and do not address the
immediate discomfort of riders.
Group 3 (Mitigation) considered many possible mitigation strategies, but
believes that only a few are achievable or feasible within the financial,
regulatory, maintenance, or operational constraints that inhibit the
actions of Valley Metro. For instance, regulations addressing sightlines and
visibility of citizen driving cars interfere with the ability of Valley Metro to
plant trees or shrubs at many stops. Further, Valley Metro may not be able
to install plants or devices that use water like water fountains or misters at
light rail stops because of adjacent high voltage electricity lines. In
addition, many strategies are not feasible due to budget or maintenance
constraints.
Using the categories defined by Connors et al. (2013), Group 3
(Mitigation) categorized mitigation strategies most relevant to the users of
Valley Metro transit system in Tempe. The unit costs of chosen strategies
summarized in the third column below and the benefits of the strategies
are addressed in the next section.
Table 14 – Urban Heat Island Causes and Mitigation Strategies
Intra-Urban Variation

Mitigation Strategies

Unit Cost or Notes

Urban fabric

Cool pavement

$2~2.5 per sqft

Urban fabric

Green pavement

$1.5~5.74 per sqft

Urban cover

Cool roofs

Not priced

Urban cover

Green roofs

$10 per sqft

Page 68 of 94

Urban cover

Green walls

$90~135 per sqft

Urban cover

Solar panels

Material: $2.87~3.85 per
watt; Labor: $0.44 per watt

Urban cover

Grass

Material: $0.05 per sqft;
Labor: $ 30 per hour

Urban cover

Shrubs

Material: $10~16.65 per
gallon; Planting: $106~2423
per tree

Urban cover

Trees

Material: $7.58~28 per
gallon; Planting: $106~2423
per trees

Urban cover

Parks

Not applicable

Urban cover

Water detention ponds

Not applicable

Urban cover

Wetlands

Not applicable

Urban metabolism

Air conditioning

Not feasible now, but may
be in the future as
temperatures increase

Urban metabolism

Fans

Not feasible now, but may
be in the future as
temperatures increase

Urban metabolism

Misting systems

Not feasible now, but may
be in the future as
temperatures increase.
Cannot be installed at light
rail stops.

Urban metabolism

Water fountains / coolers

Material: $60~180 each
Labor: $320~520 each
Excavation: $0.047~2.28 per
sqft

Urban metabolism

Public chilled water system

Page 69 of 94

Not relevant

Urban metabolism

Cooling centers

Not subject to unit pricing

Urban metabolism

Altered transit schedules in
response to extreme heat
events

Variable

Urban metabolism

Combined cooling, heat,
and power or co-generation
systems for transit centers

Depends on the needs
assessment and size

Urban metabolism

Community-based outreach
program

Depends on size and type
of program

Urban metabolism

Enhanced transit
connectivity

Depends on scale and
scope

Urban metabolism

Hot weather warning systems

Variable

Urban metabolism

Improved transit maps

No additional costs
anticipated

Urban metabolism

Intelligent signage

Variable

Urban metabolism

Pedestrian amenities

Depends on type of
amenity

Urban metabolism

Restructuring traffic and
transportation

Unknown

Urban metabolism

Transit apps

Valley Metro is working on

Urban metabolism

Using renewable energy and
energy saving appliances
and equipment

Not applicable

Urban structure

Complete Streets

Not applicable

Page 70 of 94

3.3.4 COSTS OF RECOMMENDED HEAT MITIGATION STRATEGIES
The most cost effective and sustainable strategy to mitigate heat
exposure associated with users of the Valley Metro transit system is
vegetation – especially trees. Trees have many advantages. They
reduce urban heat island and absorb greenhouse gasses. They not only
improve the aesthetic experience of the urban environment but also
contribute to both physical and psychological health. Other strategies
include:


Solar panels, which provide shade and clean energy, which could
be potentially used to run air conditioners or fans at the transit stops;



Cool pavement, which reflects solar radiation and absorbs less heat
while requiring less maintenance.



Green roofs and green pavements that reduce temperatures and
improve comfort.



Water fountains that allow transit riders to stay hydrated, but may
be prohibited at light rail stops to the presence of high voltage
electrical wires and concerns about homeless people taking “bird
baths.”



Of the strategies that might help alleviate the suffering of transit
riders, green walls and the water fountains are the most expensive.



The most cost effective are cool and green pavement.

Unit costs for mitigation strategies that Group 3 (Mitigation) believes are
feasible for transit stops operated by Valley Metro are included in Table 15
below under the categories of Urban Cover, Urban Fabric and Urban
Metabolism from Table 14 above. Costs are subject to change depending
upon the type of materials used and methods employed for installation.

Page 71 of 94

As an example of a possible mitigation strategy that might be effective at
transit stops, a member of Group 3 (Mitigation) measured the footprint of
the UDCA stop at six feet by ten feet. One possible way to reduce the
exposure of riders to heat would be to install green roofs on shelters.
According to the Table 15 below, green roof costs about $10 per square
foot, or about $600 for a 60-square foot roof. Therefore, the cost of
materials would be relatively modest but the lifecycle cost of maintaining
green roofs on many transit stops would be substantial and perhaps
prohibitive. Current bus stop shelters could not accommodate green
roofs. For an idea of how much Valley Metro would be willing to pay for
shelters, the 2008 Bus Stop Design Guidelines shows allowances for bus
stops in four levels with associated components. At the low end, the cost
of a bus stop sign is limited to $150. At the high end, the cost for an
enhanced Level 4 shelter is capped at is $23,200 (Bus Stop Design
Guidelines, p. 9).
Table 15 - Costs of Feasible Mitigation Strategies

Page 72 of 94

Descriptions of benefits associated with the selected mitigation strategies
follow.
3.3.4.1 VEGETATION
The temperature of urban surface areas may be decreased through
evapotranspiration by planting trees and vegetation. Shade provided by
trees and vegetation in public spaces mitigate the effect of heat
exposure. In addition, trees and vegetation reduce storm water runoff
and protect soil from erosion. Group 3 (Mitigation) gathered data on costs
of from multiple sources, including vendor websites and technical reports.
Group 3 (Mitigation) included only cost data for shade trees and plants
that can survive Arizona’s harsh weather conditions.
3.3.4.2 COOL AND GREEN PAVEMENT
Cool pavements reflect solar radiation, lower the surface temperatures,
and reduce the amount of heat absorbed into the pavement so that they
stay cooler than the traditional pavements. Cool pavements use
reflective aggregate, reflective clear binder, or a reflective surface
coatings. The many benefits include energy savings, emission reductions,
improved comfort and health, increased driver safety, improved air
quality, reduced street lighting cost, reduced power plant emissions,
improved water quality and slowed climate change. The properties of
solar reflectance, thermal emittance, and permeability play significant
roles. If pavement reflectance throughout a city were increased from 10
percent to 35 percent, the air temperature could potentially be reduced
by 1 degree Fahrenheit ºF (0.6 degrees Celsius (Pomerantz, Pon, Akbari, &
Chang, 2000). There are many types of cool pavements including:


Conventional asphalt or concrete;



Vegetated and non-vegetated permeable; and

Page 73 of 94



Other reflective types like resin based and colored asphalt and
colored concrete.

Costs depend on the region, climate, contractor, time of year,
accessibility of a site, underlying soils, project size, expected traffic, and
desired the life of the pavement (Ferguson et al., 2008). The benefits of
cool pavements include:


Increased water percolation, improved stormwater management
and reduced water runoff due to permeability (Booth & Leavitt,
1999);



Decreased air temperatures thereby enhancing outdoor comfort
and public health (Pomerantz et al., 2000);



Increased energy savings, reduced formation of ground-level
ozone, and reduced greenhouse gases;



Increased driver safety and reduced need for street lighting due to
improved solar reflectance



Increased durability due absorbing less solar energy over time



Less damage to local watersheds due to solar reflectance and
permeability (James & Shahin, 1998)



Mitigation of urban heat island (Qin, 2015)



Sunlight reflectance: Instead of absorbing 80~90% of sunlight as
tradition pavement does and heating the material to temperature
between 120-150°F (48–67°C), cool pavement stays cooler because
it reflects sunlight.



Reduced storm water runoff: Cool pavement is permeable, which
allows water to soak into the ground, reducing runoff and filtering
pollutants and prevent thermal shock to plants.

Page 74 of 94



Energy saving from air conditioning: Cool pavement lowers the
outdoor temperature, leading to reduce energy bills for air
conditioning needed to cool buildings.



Reduced street light cost: Cool pavement could increase the solar
reflectance of roads, reducing the energy required for street
lighting at night.



Reduced power plant emissions: Reduced energy demand reduces
emissions from power plant operation.



Improved air quality: Cool pavement reduces outside air
temperature, improving the comfort and reducing heat-related
illness. Cooler temperature could reduce smog, which is formed by
the reaction of sunlight and carbon-hydrocarbons and nitrogen
oxides in the air and could affect people’s respirations.



Improved local comfort: Cool pavements in parking lots or other
areas where people congregate or children play can provide a
more comfortable environment.



Slowed climate change: Cool pavement could reduce earth
surface temperature and consequently air temperature due to solar
reflectance, which may offset the global warming caused by
greenhouse gas in atmosphere.

3.3.4.3 GREEN WALLS
Green walls are covered with vegetation usually supported by an
integrated water delivery system. They keep the wall comparatively
cooler especially when exposed to heat or sun. They can be of great help
in the urban environments and improve the aesthetics of the area. Green
walls can be used at the transit stops to prevent direct heat exposure and

Page 75 of 94

keep the waiting area cooler from all the sides caused by
evapotranspiration and reflective properties.
3.3.4.4 GREEN ROOFS
Green roofs or rooftop gardens reduce the heat exposure of the
occupants. They not only help in reduce heat exposure but also
contribute to improved human health. They help to reduce the urban
heat island effect and reduce the greenhouse gases. They can be
efficiently applied at the transit stops to reduce the temperature
underneath.
3.3.4.5 SOLAR PANELS
Solar panels can be used to generate electricity and provide shade. The
electricity generated through the panels can be used to power lights,
fans, and cell phone chargers at transit stops. Solar panels are especially
attractive in Phoenix because of the abundance of sunlight.
3.3.4.6 WATER FOUNTAINS
The human body depends on water to survive. Every cell, tissue and organ
in the body needs water to work properly. The body uses water to
maintain the temperature, remove waste, and lubricate the joints.
Dehydration is a danger for people living and working outdoors in Arizona
during the summer. Widely installed water fountains could increase access
to the water at transit stops, where people are vulnerable to the heat
while waiting. Drinking water is the most effective way to reduce body
temperature and to prevent people from suffering from heat-related
illness or death
3.3.4.7 FUTURE INNOVATION
When exploring current possibilities, it is easy to leave out what may be
possible in the future. For instance, in the future it is possible that transit

Page 76 of 94

entities will invent new ways to communicate with users so that transit
systems can quickly transition from passively protecting users to actively
moving people to cooler and safer locations.
One strategy that has already been implemented by some transit
agencies like SFMTA, King Country Metro, but not all is the transit
frequency map. Transit consulting planner Jarrett Walker (2010) proposed
this style of map as a response to what he perceived as misleading transit
maps. By giving all lines equal weighting on the system-wide map, it is
difficult to see which buses run less frequently than others. He asserts that
the way the transit system looks to the layperson is the most important
factor in the system.
In Phoenix, this innovation might be even more critical. By giving all lines
equal weights, it is difficult to know what kind of service is being run. For
instance, on the current (April 2017) map, the graphic representation for
Route 72, a bus line that has headways of 10 minutes at peak hours is
given the same representation as Route 251, which runs at an hour
frequency throughout the week. For casual transit users, this can be
misleading, and might encourage waiting for a bus that might be
infrequent despite the fact that the width of the route on the map shows
graphically that it has the same frequency.
3.3.5 FISCAL POLICY (FINDINGS AND RECOMMENDATIONS)
Most people do not think of a policy as being part of a city’s
infrastructure. However, almost all physical forms of infrastructure in a city
are results of policies and codes. In fact, the built environment is a direct
manifestation of various local, state, and other municipal policies,
regulations, and codes. While policies usually have good intentions they
also may carry unintended consequences that may negatively impact a
person's experience with urban infrastructure. For example, a building

Page 77 of 94

setback may make sense for safety and visibility reasons for automobiles,
but might unintentionally reduce the amount of shade provided to transit
users by buildings when they are built right next to lot lines. It is important
to identify policies that may be barriers to adapting to climate change
and to design new policies in ways that are flexible and easily adaptable
to outcome vulnerabilities associated the climate change.
The goal of Group 3 (Mitigation) was to investigate and understand the
policies most relevant to the creation of the built environment within the
study area as well as identifying possible policy proposals in order to
improve people’s experience with infrastructure related to heat exposure
in the future.
Group 3 (Mitigation) understands the complex nature of the policies and
regulations that have produced the infrastructures. We followed a threepronged approach to achieving our goal. First, we identified the most
relevant policies and governing bodies that create some of the conditions
that are reported by Group 1 (Transport) and Group 2 (Exposure). Second,
Group 3 (Mitigation) reviewed these policies and made necessary
updated proposals. Third, Group 3 (Mitigation) devised additional policies
to help mitigate major infrastructure deficiencies identified by the other
groups on the project team.
Finding: City of Tempe has codified shade infrastructure in two discrete
ways. First, they have used the General Plan as the major touchstone for
the priorities of the city and this has noticeably changed over the last
iteration of the plan. The 2030 version, enacted by the City of Tempe in
2003, did not mention shade or shade mitigation. The 2040 version
enacted in 2014 mentioned shade 21 times. The primary framing behind
improving the shade of the city is pedestrian friendliness in a traditional
New Urbanist/Complete streets model. There are no noticeable SMART

Page 78 of 94

strategies, which is understandable from the perspective of quantitative
evidence.
Recommendation: While it is encouraging that the City of Tempe now
acknowledges the importance of including shade in general and zoning
plans, additional improvements can be made. The 2040 version of the
General Plan includes shade from a quality of life perspective. While this is
not necessarily a deficiency it is important that future iterations of the
general plan acknowledge the important role that shade plays in
reducing personal heat exposure during extreme heat events.
Group 3 (Mitigation) recommends that the Tempe start including shading
requirements for developments that are adjacent to bus stops. This will
help reduce the cost to the city for provide shade at bus stops and will
perhaps help this strategy be implemented quicker than if it were up to
just the city to get it done through its own funding sources. Similar
requirements to this have been proven to be effective. A local example is
the City of Phoenix’s Walkable Urban Code that requires 75% shade cover
over sidewalks.
Finding: No Boards & Committees exist that exclusively deal with heat
although the Sustainability Commission & Transportation Commission
might interested in the topic.
Recommendation: First, the Sustainability and Transportation Commissions
should each be introduced to the concepts of individual heat exposure
as a result of interactions with infrastructure during extreme heat
events. In the short term, there should be cooperation between these
two boards on tackling these issues. While both of these boards should
always keep heat exposure in mind when considering issues within the
city, our long-term proposal is that the City creates a new commission
explicitly focused on the issue of heat exposure. We feel that the issue is

Page 79 of 94

important enough to warrant its own commission. There should be
cooperation between the existing two commissions and the future
commission when tackling these issues.
Finding: Maricopa County is in a unique position as the governing body
directly above most of the municipalities in the Phoenix region. Maricopa
County has influence and money available that could be used to help
guide and direct municipalities by funding studies, creating best practice
guides, and further assisting cities with heat mitigation. However,
Maricopa County does not currently provide much support for heat
mitigation.
Recommendation: Maricopa County needs to take advantage of its
hierarchal position over the local municipal governments and take
charge on issues related to personal heat exposure. This includes
providing municipalities with best practice solutions, providing a forum for
cities to exchange resources and ideas, and providing funding
opportunities for cities to take advantage of to speed up the
implementation process.
Finding: In 2015 Maricopa County released a comprehensive MultiJurisdictional Hazard Mitigation Plan. The plan focused on environmental
and climate change hazards. While extreme heat is one of the hazards
mentioned in the plan, there was not nearly as much focus on it as would
be expected from city built in the desert. The plan lays out the following
four heat-related goals:


Identify, stock, and communicate locations within the community
that can serve as cooling stations during times of extreme heat.



Perform a public information campaign at the onset of the extreme
heat season to help educate the public on ways to remain safe
during periods of extreme heat.
Page 80 of 94



Partner with non-governmental organizations like The Salvation
Army, church organizations, and homeless shelters to provide
respite care and hydration stations during extreme temperature
events.



Investigate and develop an implementation strategy for using “cool
roofs” on any new or major roof rehabilitation projects of
tribal/county/city/town owned buildings to lower the urban heat
island effects.

Recommendation: While these goals are all good, it is a very basic
framework. It can be used as a starting point but by no means should this
be used as the four-point plan moving forward. We recommend that the
county expand upon this list, providing more specific goals. These goals
should be:


Upgrading bus stops until 90% of all stops in the county meet a set
shade standard;



Identify and map all cooling stations in the county;



Create funding opportunities specifically for personal heat exposure
mitigation.

Finding: There are no grants offered by the City of Tempe, Maricopa
County, Maricopa Association of Governments, or the State of Arizona
that directly relate to heat mitigation strategies. The Federal Transit
Authority offers grants that could potentially be used for heat mitigation
strategies at bus stops. The most relevant of these offered grants is the
Buses and Bus Facilities Grant No. 5339. The function of this grant is to
provide funds to municipalities to improve buses and related facilities,
including bus stops. While this could be used to mitigate heat exposure at
bus stops, that is not its focus. In 2016, the FTA awarded 63 grants under
the program totally about $211 million. Out of these grants, only three of
Page 81 of 94

them went to projects other than bus fleet upgrades. Two of the projects
were upgrading transit centers and one was to update a bus repair
shop. None of the grants were used to upgrade bus stops or to mitigate
heat exposure in any way. Regardless, this program still appears to be
one of the better funding mechanisms for heat exposure mitigation
projects.
The lack of specific funding mechanisms is especially telling for several
reasons. First, it reflects the general lag in combating heat exposure, even
in hotter regions in the United States. There are plenty of funding
opportunities for climate change mitigation strategies but heat has not
been included in the discussions. It is important to implement mitigation
strategies because many cities in Maricopa County have acknowledged
that personal heat exposure is of concern but lack the funds to implement
effective mitigation strategies.
Recommendation: New methods of funding must be found to implement
the mitigation strategies. Traditionally, funding systems are biased in favor
of initial investment in new infrastructure projects. Maintenance costs are
usually unpredictable and therefore are not emphasized. It is also worth
noting that given the change in climate, conditions in Phoenix will
become more extreme, so the same should be assumed of maintenance.
For instance, felled trees may change monsoon patterns change. The
ability of Valley Metro to plant storm resistant trees and increase the
resilience of trees is a good method to reduce the costs of maintenance.
The cost of climate change in relation to personal heat exposure to Valley
Metro is also liable to rise. As the number of heat related health incidents
rise, litigants will target Valley Metro. More information is needed to
conduct a proper cost-benefit analysis, but Group 3 (Mitigation) expects
litigation to increase.

Page 82 of 94

3.4 LIMITATIONS
The scope of this study was limited to the area between Rio Salado Pkwy,
S. Rural Rd, E.Apache Blvd, and S Priest Dr This area contains ASU Tempe
Campus and Downtown Tempe which means that results of this study are
highly influenced by the student population and cannot be generalized
for the whole city. Additionally, the survey conducted during this research
is a small sample size of the Tempe population, and was skewed towards
the younger population (likely students as indicated above). Also, we
surveyed transit riders in the month of April which is significantly cooler
than summer in the study area (compare 85°F to 110°F in summer).
There are some limitations associated with data availability and resolution.
Daily ridership data per transit stop was not available (outside averages
for each season) to conduct sensitivity analyses between ridership and
extreme heat days. Without daily, weekly, or possibly even monthly
ridership data, it is hard to tease out the relationship between historical
temperature and ridership behavior. Transit authorities that wish to identify
how extreme heat may influence ridership should start by comparing
historical temperature data with highly resolute ridership data.
The wait time data used (Figure 6) was estimated based on Valley Metro
service frequency and does not reflect actual arrival times and schedule
delays. Therefore, actual rider wait times likely vary day-to-day from the
presented averages.
This study did not include socio-economic characteristics as they were
deemed outside the scope of the analysis. Future work would benefit
taking into account where vulnerable populations are located in a study
region to identify not only characteristics of transit service that may be
vulnerable, but also the dynamic of vulnerable riders who interact with
vulnerable transit service characteristics.

Page 83 of 94

3.5 CONCLUSION
Transportation infrastructures will be challenged in the future by more
extreme heat events, increased overall temperatures, and more riders
due to increasing populations. Extreme heat exposure is both a health risk
for riders and threat to the operations of Valley Metro. To increase
resilience of its transit system, Valley Metro must deploy robust mitigation
strategies that decrease outcome vulnerability caused climate change
and prepare to mitigate the consequences of more extreme heat events
that will damage transportation infrastructures. At the same time, all of
levels of government should begin to address contextual vulnerability at
risk groups of riders like the elderly and homeless who have no choice but
to use transit regardless of the heat exposure.

Page 84 of 94

4 REFERENCES:
Al-musaed, A. (2007, September). Evaporative cooling process adaptive
for Baghdad city climate. In 352nd PALENC conference and 28th
AIVC conference on building low energy cooling and advanced
ventilation technologies in the 21st century. Booth, D. B., & Leavitt, J.
(1999). Field evaluation of permeable pavement systems for
improved stormwater management. Journal of the American
Planning Association, 65(3), 314-325.
Arizona Living Landscape & Design. (2015). “Arizona Desert Plants”. <
http://landscapearizona.com/blog/category/arizona-desert-plants/
>. (May 1, 2017)
Arizona State University. (2017). “ASU Solar.” < https://cfo.asu.edu/solar>
(May 1, 2017)
ASU College of Design. No date. OASIS: Oriented Accessible Street
Integrated Shelter: Bus Stop Shading Study and Design Services.
AZDHS, 2015. Heat Caused & Heat-Associated Deaths in Arizona by Year
(2003-2014) [WWW Document]. URL
http://azdhs.gov/documents/preparedness/epidemiology-diseasecontrol/extreme-weather/pubs/heat-related-deaths-updated-may2015.pdf (accessed 1.1.17).
Berko, J., Ingram, D.D., Saha, S., Parker, J.D., 2014. Natl. Health Stat.
Report. 1–15.
Bois, P., D.L. Childers, T. Corlouer, J. Laurent, A. Massicot, C.A. Sanchez, A.
Wanko. 2017. Confirming a Plant-Mediated Biological Tide in an
Aridland Constructed Treatment Wetland. Ecosphere.
Boukhabl, M., & Alkam, D. (2012). Impact of vegetation on thermal
conditions outside, Thermal modeling of urban microclimate, Case
study: the street of the republic, Biskra. Energy Procedia, 18, 73-84.
Bus Stop Design Guidelines. (2008). Regional Public Transportation

Page 85 of 94

Authority. Retrieved from
http://www.valleymetro.org/images/uploads/projects/Bus_Stop__Design_Guidelines.pdf
CDC, 2006. Morb. Mortal. Wkly. Rep. 55, 796–798.
Chuang, W.-C., Karner, A., Selover, N., Hondula, D., Chhetri, N., Middel,
A., Roach, M., Dufour, B., 2015. Arizona Extreme Weather, Climate
and Health Profile Report.
City of Tempe, 2016. Transportation Master Plan.
Comprehensive R Archive Network. 2017. R Statistical Package. Retrieved
from: https://cran.r-project.org/
Connors, J. P., Galletti, C. S., & Chow, W. T. (2013). Landscape
configuration and urban heat island effects: assessing the relationship
between landscape characteristics and land surface temperature in
Phoenix, Arizona. Landscape ecology, 28(2), 271-283.
Cool Pavements (2017). Cool Pavements | HEATISLAND. Retrieved from
https://heatisland.lbl.gov/coolscience/cool-pavements
Eisenman, D. P., Wilhalme, H., Tseng, C. H., Chester, M., English, P., Pincetl,
S., ... & Dhaliwal, S. K. (2016). Heat death associations with the built
environment, social vulnerability and their interactions with rising
temperature. Health & place, 41, 89-99.
Eisenman, D.P., Wilhalme, H., Tseng, C., Chester, M., English, P., Pincetl, S.,
Fraser, A., Vangala, S., Dhaliwal, S.K., 2016. Health Place 41, 89–99.
Energysage. (2017). “How much do solar panels cost in the U.S.?” <
http://news.energysage.com/how-much-does-the-average-solarpanel-installation-cost-in-the-u-s/>. (May 1, 2017)
Ennos, R. 2015. Can trees really cool our cities down? The Conversation.
Retrieved from http://theconversation.com/can-trees-really-cool-ourcities-down-44099
Environment Protection Agency (EPA). (2016). “Heat Island Effect.” <

Page 86 of 94

https://www.epa.gov/heat-islands/heat-island-cooling-strategies>
(May 1, 2017)
Familydoctor.org editorial staff. (2017). “Hydration: Why It’s So Important”
< https://familydoctor.org/hydration-why-its-so-important/> (May 1,
2017)
Fan, Y., Guthrie, A., Levinson, D., 2016. Transp. Res. Part A Policy Pract. 88,
251–264.
Ferguson, B., Fisher, K., Golden, J., Hair, L., Haselbach, L., Hitchcock, D., ...
& Waye, D. (2008). Reducing Urban Heat Islands: Compendium of
Strategies-Cool Pavements.
FIXr. (2015). “Fountain Installation Cost”. <
https://www.fixr.com/costs/fountain-installation > (May 1, 2017)
Fraser, A.M., Chester, M. V., 2016. J. Transp. Heal.
Goldman, T., Corbett, S., & Wachs, M. (2001). Local Option Transportation
Taxes in the United States (Part One). Institute of Transportation
Studies.
Hajat S, Armstrong B, Baccini M, Biggeri A, Bisanti L, Russo A, Paldy A,
Menne B and Kosatsky T 2006 Impact of high temperatures on
mortality: is there an added heat wave effect? Epidemiology 17 6328
Harlan SL, Declet-Barreto JH, Stefanov WL, Petitti DB, 2013. Neighborhood
Effects on Heat Deaths: social and Environmental Predictors of
Vulnerability in Maricopa County, Arizona. Environmental Health
Perspectives 121(2), 197-204.
Harlan, S.L., Declet-barreto, J.H., Stefanov, W.L., Petitti, D.B., 2013. Environ.
Health Perspect. 121, 197–204.
Hartmann DL, Klein Tank AMG, Rusticucci M, Alexander LV, Brönnimann S,
Charabi Y, et al. 2013. Observations: atmosphere and surface. In:
Stocker TF, Qin D, Plattner G-K, Tignor M, Allen SK, Boschung J, Nauels

Page 87 of 94

A, Xia Y, Bex V, Midgley PM, editors. Climate change 2013: the
physical science basis. Contribution of working group I to the fifth
assessment report of the intergovernmental panel on climate
change. Cambridge: Cambridge University Press; 2013.
Haynes, M. (2017) “Re: Financial quote for green wall installation request.”
Message to Reddy, Y.S.R. April 12, 2017.
Hondula DM, RC Balling Jr., JK Vanos, M Georgescu. 2015. Rising
Temperatures, Human Health, and the Role of Adaptation. Curr Clim
Change Rep 1: 144-154
Howmuch website staff. (2016). “Tree Planting Costs”. <
https://howmuch.net/costs/tree-install>. (May 1, 2017)
James, W., & Shahin, R. (1998). A laboratory examination of pollutants
leached from four different pavements by acid rain. Advances in
modeling the management of stormwater impacts, 6(17), 321.
Jenerette, G. D., Harlan, S. L., Stefanov, W. L., & Martin, C. A. (2011).
Ecosystem services and urban heat riskscape moderation: water,
green spaces, and social inequality in Phoenix, USA. Ecological
Applications, 21(7), 2637-2651.
Karner, A., Hondula, D. M., & Vanos, J. K. (2015). Heat exposure during
non-motorized travel: Implications for transportation policy under
climate change. Journal of Transport & Health, 2(4), 451-459.
Karner, A., Hondula, D.M., Vanos, J.K., 2015. J. Transp. Heal. 2, 451–459.
Kim, H.H., 2007. Int. J. Remote Sens. 13, 2319–2336.
LiveWall.(2017). “How to Plant a Green Wall.” <
http://livewall.com/technical/planting-guide/>. (May 1, 2017)
Luber, G., McGeehin, M., 2008. Am. J. Prev. Med. 35, 429–435.
McGregor G R, Pelling M, Wolf T and Gosling S 2007 Social impacts of
heat-waves Science Report SC020061/SR6 (UK Environment Agency)
Meehl G A and Tebaldi C 2004 More intense, more frequent, and longer

Page 88 of 94

lasting heat waves in the 21st century Science 305 994-7
Middel, A., Selover, N., Hagen, B., Chhetri, N., 2016. Int. J. Biometeorol. 60,
1849–1861.
Mills, D., Schwartz, J., Lee, M. et al. Climatic Change (2015) 131: 83.
doi:10.1007/s10584-014-1154-8
Mountain Sates Wholesale Nursery. (2004). “Caesalpinia Mexicana
Mexican bird of paradise”. <
http://www.mswn.com/media/info_sheets/caesalpinia_mexicana.pd
f> (May 1, 2017)
National Research Council Transportation Research Board. 2008. Special
Report 290: Potential Impacts of Climate Change on US
Transportation Infrastructure.
O’Brien, K., Eriksen, S., Nygaard, L. P., & Schjolden, A. (2007). Why different
interpretations of vulnerability matter in climate change discourses.
Climate policy, 7(1), 73-88.
Pomerantz, M., Pon, B., Akbari, H., & Chang, S.-C. (2000). The effects of
pavements' temperatures on air temperatures in large cities.
Qin, Y. (2015). A review on the development of cool pavements to
mitigate urban heat island effect. Renewable and Sustainable Energy
Reviews, 52, 445-459.
Saha, M. V, Davis, R.E., Hondula, D.M., 2013. Am. J. Epidemiol. 179, 467–
474.
Sarofim, M.C., Saha, S., Hawkins, M.D., Mills, D.M., Hess, J., Horton, R.,
Kinney, P., Schwartz, J., St. Juliana, A., 2016. Ch. 2: TemperatureRelated Death and Illness, in: The Impacts of Climate Change on
Human Health in the United States: A Scientific Assessment. U.S.
Global Change Research Program, Washington, DC, pp. 43–68.
Tan, J., Zheng, Y., Tang, X., Guo, C., Li, L., Song, G., 2010. Int. J.
Biometeorol. 54, 75–84.

Page 89 of 94

Uejio, C.K., Wilhelmi, O. V, Golden, J.S., Mills, D.M., Gulino, S.P., Samenow,
J.P., 2011. Health Place 17, 498–507.
United States Environmental Protection Agency (EPA). “Using Cool
Pavements to Reduce Heat Islands.”
Valley Metro, 2017. Ridership Reports
Weller, N.A., D.L. Childers, L. Turnbull, and R.F. Upham. 2016. Aridland
constructed treatment wetlands I: Macrophyte productivity,
community composition, and nitrogen uptake. Ecol. Engineering.
Yamamoto, Y. (2006). Measures to mitigate urban heat islands. Science
and Technology Trends Quarterly Review, 18(1), 65-83.
Yip F Y, Flanders W D, Wolkin A, Engelthaler D, Humble W, Neri A, Lewis L,
Backer L and Rubin C 2008 The impact of excess heat events in
Maricopa County, Arizona: 2000-2005 Int. J. Biometeorol. 52765-72

Page 90 of 94

5 APPENDIX:
5.1 APPENDIX A: TEMPE TRANSIT SURVEY FOR STUDY REGION

Page 91 of 94

Page 92 of 94

5.2 APPENDIX B: STANDARDIZED CAP LTER METADATA PACKAGE
ASSOCIATED WITH THIS PROJECT DATA.
For future compatibility with CAP LTER or other database for data publication
purposes. Project data available upon request.
Attribute Name
sample_id
date
researchers
stop_id
nextride_id
location
transit_mode
timestamp

image_id

thermal_image_id

image_subject

Attribute
Description
Observation
number
Date of
observation
Researchers
present
ValleyMetro stop
ID
ValleyMetro
NextRide ID
Closest major
crossroads
Transit mode
utilizing stop
Classification of
observed bird
activity at time of
observation
Corresponding
database file
number
Corresponding
database file
number
Transit stop
component

surface_temp_f

Surface
temperature

air_temp_f

Air temperature

relative_humidity_
%
min_temp_f

Relative humidity

max_temp_f

Maximum
temperature

temp_range_f

Temperature
range

subject_shade

Shading of image
subject

Minimum
temperature

Units
integer

Unit definition/details

Other notes

Sequential

dd/mm/y
y
text
Integer
Integer
Text
Text

Bus or light rail

Text

Time in 24 hour format
based on MST

Integer

File number of normal
camera image in
thermal image database
File number of thermal
camera image in
thermal image database
Component of transit
stop infrastructure that is
subject of image

Integer

Text

Degrees
fahrenhei
t
Degrees
fahrenhei
t
%

Temperature of image
subject as captured by
FLIR camera
Temperature of air 1m
from ground at center of
transit stop

Degrees
fahrenhei
t
Degrees
fahrenhei
t
Degrees
fahrenhei
t

Minimum temperature
captured in thermal
image
Maximum temperature
captured in thermal
image
Difference between
minimum and maximum
temperatures in thermal
image
Indicates whether
subject was located in
shade

Integer

Page 93 of 94

Overview,
seats,
pavement,
grass,

0 = not
shaded, 1 =
shaded

material

Dominant
pavement
material
Amount of green
land cover

Integer

stop_shade

Stop shade
features

Integer

seating

Stop seating
features

Integer

%green_space

%

Dominant pavement
material within 3m of
transit stop
Amount of green
landcover within 3m of
transit stop
Indicates which shade
features stop has

Indicates if stop has
seating features

Page 94 of 94

1 = concrete, 2
= asphault, 3
=brick
25%
increments
0 = no shading
1 = direct
shading from
transit structure
(awning, etc.)
2 = indirect
tree shade
3 = direct tree
shade
0 = no seating
1 = seating