The Professional Geographer

ISSN: 0033-0124 (Print) 1467-9272 (Online) Journal homepage: https://www.tandfonline.com/loi/rtpg20

Vulnerability to Extreme Heat in Metropolitan
Phoenix: Spatial, Temporal, and Demographic
Dimensions
Winston T. L. Chow , Wen-Ching Chuang & Patricia Gober
To cite this article: Winston T. L. Chow , Wen-Ching Chuang & Patricia Gober (2012) Vulnerability
to Extreme Heat in Metropolitan Phoenix: Spatial, Temporal, and Demographic Dimensions, The
Professional Geographer, 64:2, 286-302, DOI: 10.1080/00330124.2011.600225
To link to this article: https://doi.org/10.1080/00330124.2011.600225

Published online: 18 Aug 2011.

Submit your article to this journal

Article views: 1996

View related articles

Citing articles: 34 View citing articles

Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=rtpg20

Vulnerability to Extreme Heat in Metropolitan Phoenix:
Spatial, Temporal, and Demographic Dimensions∗
Winston T. L. Chow, Wen-Ching Chuang, and Patricia Gober
Arizona State University
This study assessed the spatial distribution of vulnerability to extreme heat in 1990 and 2000 within metropolitan Phoenix based on an index of seven equally weighted measures of physical exposure and adaptive capacity.
These measures were derived from spatially interpolated climate, normalized differential vegetation index,
and U.S. Census data. From resulting vulnerability maps, we also analyzed population groups living in areas of
high heat vulnerability. Results revealed that landscapes of heat vulnerability changed substantially in response
to variations in physical and socioeconomic factors, with significant alterations to spatial distribution of vulnerability especially between eastern and western sectors of Phoenix. These changes worked to the detriment
of Phoenix’s Hispanic population and the elderly concentrated in urban-fringe retirement communities. Key
Words: adaptive capacity, physical exposure, urban heat island, vulnerability.

Este estudio evaluó la distribución espacial de la vulnerabilidad al calor extremo en 1990 y el 2000 dentro
del área metropolitana de Phoenix, sobre la base de un ı́ndice de siete medidas igualmente ponderadas de
exposición fı́sica y capacidad de adaptación. Estas medidas se derivan del clima interpolado espacialmente, del
ı́ndice normalizado de vegetación diferencial, y datos censales de EE.UU. A partir de mapas de vulnerabilidad
también se analizaron grupos de población que viven en zonas con vulnerabilidad a las altas temperaturas.
Los resultados revelaron que los paisajes con vulnerabilidad al calor cambiaron sustancialmente en respuesta a
variaciones en factores fı́sicos y socioeconómicos, con modificaciones importantes en la distribución espacial
de la vulnerabilidad, especialmente entre los sectores este y oeste de Phoenix. Estos cambios se dieron en
detrimento de la población hispana de Phoenix y los ancianos concentrados en comunidades de jubilación
urbano-marginales. Palabras claves: capacidad de adaptación, exposición fı́sica, isla de calor urbano,
vulnerabilidad.

M

etropolitan Phoenix, encompassing the
City of Phoenix and twenty-six other
municipalities and Native American communities in central Arizona (Figure 1), is one of
the nation’s largest and fastest growing urban agglomerations in both population size
and land area (U.S. Census 2007). Its subtrop-

ical desert location presents an extreme climate with low total annual precipitation and
high average maximum temperatures. Exposure to extreme heat is a potentially serious
issue to its residents, especially during the summer months when daytime temperatures regularly exceed 43◦ C (110◦ F). Extreme heat is

∗This article was developed from a poster presented at the 2009 Association of American Geographers Annual Meeting in Las Vegas, Nevada. We

thank Anthony Brazel, Darren Ruddell, Nancy Selover, Chona Sister, and Sally Wittlinger (Arizona State University) for their helpful comments.
We also appreciated critiques from three anonymous referees and the editor that considerably improved this article. This material is based on
work supported by the National Science Foundation under Grant SES-0345945, Decision Center for a Desert City. Any opinions, findings, and
conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science
Foundation.
C Copyright 2012 by Association of American Geographers.
The Professional Geographer, 64(2) 2012, pages 286–302 
Initial submission, October 2009; revised submissions, May and August 2010; final acceptance, September 2010.
Published by Taylor & Francis Group, LLC.

Vulnerability to Extreme Heat in Metropolitan Phoenix 287

Figure 1 Metropolitan Phoenix and its constituent cities and census-designated places in 2000. (Color
figure available online.)

exacerbated by the growth of the city’s strong
near-surface urban heat island (UHI), which
can raise nighttime temperatures by more than
6◦ C (10.8◦ F; Baker et al. 2002). Prolonged exposure to high temperatures has been shown to
be hazardous to human health, especially when
critical temperature thresholds are suddenly exceeded as, for example, during early onset of
summer (Sheridan and Kalkstein 2004). Severe
temperatures also have detrimental effects on
human morbidity and mortality, as seen in the
July 1995 Chicago heat wave that killed 700
(Semenza et al. 1996; Klinenberg 1999) and in
the July and August 2003 Western European
event that killed 35,000 (Stott, Stone, and Allen
2004). More than 250 deaths in Arizona were
attributable to exposure to extreme heat between 1993 and 2002; the rate of age-adjusted
deaths from heat exposure was the highest in

the nation (Centers for Disease Control and
Prevention [CDC] 2005).
The purpose of this article is to examine the
spatial and temporal changes in vulnerability
to extreme heat in Phoenix between 1990 and
2000. To this end, we (1) developed a metric for heat stress vulnerability based on measures of physical exposure (e.g., maximum and
minimum temperatures) and adaptive capacity
of the human population (e.g., income and social stability), (2) examined the spatial pattern
of this metric in both 1990 and 2000, and (3)
determined what proportion of different racial
and ethnic minority populations lived above a
critical threshold in vulnerability in both time
periods. We refer to Phoenix as the entire
metropolitan area and City of Phoenix as the
municipality that constitutes the metropolitan
core city.

288 Volume 64, Number 2, May 2012
Urban Heat Island Research in
Phoenix
The UHI is the most obvious, yet inadvertent, impact of urbanization on local-scale
weather and climate. It is generally strongest
at night when excess heat stored in urban
surfaces during the daytime is released to the
atmosphere. Its intensity is determined by
factors such as surface material type, sky-view
factor (exposure of urban surfaces to the sky),
and synoptic weather conditions (Oke 1982).
The large-scale urbanization of Phoenix has
been accompanied by a UHI of growing
intensity and expanse (Hsu 1984; Sun et al.
2009). Maximum magnitudes occur during
summer nighttime hours and can exceed 8◦ C
under ideal conditions (Balling and Cerveny
1987; Hedquist and Brazel 2006). Urban–rural
temperature differences are either negligible
(∼1–2◦ C) during the day or lower in the
city where irrigated residential landscaping
and surrounding farmlands increase evapotranspiration, resulting in a cooling effect—a
phenomenon called the “desert oasis effect”
(Brazel et al. 2000). At night, temperatures are
highest in the urban core and diminish with
proximity to the urban fringe (Balling and
Brazel 1987). UHI intensity also increases in
response to land-use change from new home
construction and land development (Hawkins
et al. 2004; Grossman-Clarke et al. 2005; Brazel
et al. 2007; Grossman-Clarke et al. 2008).
A parallel set of studies addressed the
human consequences of UHI development in
Phoenix, including increased “misery hours
per day” where apparent temperatures exceed
37◦ C (100◦ F) and greater cooling degree hours
per year (Baker et al. 2002). The UHI has
also been shown to increase residential water
use (Guhathakurta and Gober 2007), costs
of cooling buildings during peak energy use
summer months (Golden et al. 2008), and the
increased potential for heat stress, especially
among vulnerable segments of the urban
population (Harlan et al. 2006).
Given these detrimental impacts, more attention is now being paid to physical and social
indicators of vulnerability to extreme heat in
Phoenix and elsewhere. These indicators can
potentially guide mitigation strategies. These
include replacing impervious surfaces with
more porous materials (Stone and Norman

2006), modifying thermal characteristics of
buildings by increasing albedo (Sailor 1995),
and using materials with low thermal inertia
for individual buildings (Emmanuel and
Fernando 2007). A particularly fruitful UHI
mitigation strategy in Phoenix is to increase
the spatial coverage of irrigated vegetated
surfaces (Gober et al. 2010), which has been
shown to be effective in reducing near-surface
urban temperatures in several City of Phoenix
suburban neighborhoods (Stabler, Martin, and
Brazel 2005).

Assessing Vulnerability to Heat
Stress
Vulnerability of human populations to extreme
temperatures and other environmental hazards
is usually defined as the degree to which they
are likely to experience harm due to exposure
(Turner et al. 2003). Such harm depends on
both (1) physical exposure to extreme heat and
(2) a population’s adaptive capacity. The latter
is the ability to mitigate risk through mechanisms such as air-conditioning or irrigated
landscaping to cool areas surrounding the home
and immediate neighborhood or through adjustments in behavior, such as staying indoors
during excessive heat warnings or finding help
in case of an emergency instead of remaining
socially isolated. Conceptualizations of human
vulnerability to any environmental risk thus require attention to the physical event itself and
to social vulnerability, the capacity of the social
system to respond and adapt to that risk (Cutter
and Finch 2008).
The capacity to respond to hazards has been
linked to racial and ethnic status, income level,
gender, age (young children and the elderly),
migration status, and housing tenure (Ngo
2001; Heinz Center 2002; Wisner et al. 2004;
Mayhorn 2005; National Research Council
2006). Populations lacking in economic assets
and access to public support systems, with
diminished physical or cognitive capacities to
respond to warnings and missing strong and
enduring social support systems, are least able
to adapt and thus the most vulnerable to a
hazardous event.
Attempts to map vulnerability are fraught
with data comparability and methodological and conceptual challenges. The onset of

Vulnerability to Extreme Heat in Metropolitan Phoenix 289
climate change has increased the imperative to
map and explain social vulnerability at national
and subnational levels and changes therein. In
a U.S. county-level study, Cutter and Finch
(2008) found that the quality of the built environment, age, race and ethnicity, and gender
accounted for half of the spatial variation in
social vulnerability in U.S. counties. Substantial changes in distribution occurred between
1960 and 2000, with high vulnerability becoming more focused on the U.S.–Mexico border region. High social vulnerability in 2000
was associated with urban development, race
and ethnicity, and poverty. At the subnational
scale, Cutter, Mitchell, and Scott (2000) included both biophysical and social indicators
into an assessment of vulnerability in Georgetown County, South Carolina. They found that
areas of the highest biophysical vulnerability
did not overlap with areas of the highest social vulnerability. The most vulnerable places
instead combined medium levels of biophysical vulnerability with medium to high levels
of social vulnerability. The significance of this
finding is that it would take only a moderatelevel physical event to disrupt the livelihoods
and well-being of many residents.
There have been numerous efforts to assess vulnerability to hazards in Phoenix from
an environmental equity and social justice perspective (Bolin et al. 2000; Bolin et al. 2002;
Grineski, Bolin, and Boone 2007). Bolin et al.
(2000), for example, studied the population
characteristics of those exposed to hazardous
industrial and toxic waste sites and found that
these were disproportionately located in areas
where low-income, racial and ethnic minorities
tend to live. The spatial concentration of hazardous facilities, in conjunction with the segregated nature of the City of Phoenix’s disadvantaged populations, led to a highly unequal
environmental burden borne by the region’s
lower income and racial minority populations.
The problem of heat stress as a hazard has
gained currency in Phoenix in light of the intensifying UHI and in anticipation of a warmer climate associated with climate change. Harlan et
al. (2006) recorded temperatures in eight City
of Phoenix neighborhoods of differing socioeconomic characteristics and developed a human thermal comfort index (a measure of heat
stress) based on the energy balance of a person exposed to the surrounding microclimate

and thermal radiative environment. High heat
stress exposure was significantly and positively
correlated with high population densities and
heavily Hispanic populations and negatively associated with access to open spaces and irrigated vegetation and income. Homes in areas
with higher-than-average physical exposure to
heat were, in the main, less well adapted to accommodate heat stress (i.e., no swimming pools
and lower albedo roofs), as well as less welldeveloped social networks. Equally troubling
was that many poor neighborhoods exposed to
severe heat housed residents who spoke only
Spanish and who were newcomers to the city.
Thus, high physical exposure was coincident
with high social vulnerability, rendering poor
residents highly vulnerable to harm from extreme heat.
Given the importance of irrigated vegetation
in UHI mitigation, Jenerette et al. (2007) used
a path model to examine social determinants of
surface temperature and vegetation patterns.
They argued that social vulnerability to heat
operates through the ability to (1) modify land
cover by vegetation and (2) live at lower urban
densities. In their view, well-off Phoenicians
used superior social and economic status to
maintain low-density housing units with much
irrigated vegetation to reduce heat stress.
Golden et al. (2008) documented deleterious
health consequences of heat exposure using
information about 2001 through 2006 heatrelated emergency dispatches (HRD) for most
of Phoenix. Annual, monthly, and day-of-week
distributions of HRD were correlated with
several climate variables, including a human
comfort index based on assorted climate
data from several Phoenix climate stations.
Heat-related health emergencies were strongly
related to maximum temperature, elevated
human comfort and heat indexes, and temporal
exposure to excessively high solar irradiance,
especially during summer.
Maximum temperature is one of the most
important components in the physical exposure
to heat vulnerability in Phoenix. Ruddell et al.
(2009) developed a “riskscape” of heat stress
across forty City of Phoenix neighborhoods
during a four-day heat wave in July 2005 and
found that (1) the distribution of extreme heat
varied in space; (2) neighborhoods exposed
to higher temperatures significantly correlated with results of survey respondents who

290 Volume 64, Number 2, May 2012
perceived greater heat stress; and (3) elderly,
minority, and low-income residents were more
exposed to heat stresses than their younger,
white, more affluent counterparts.
The substantial and growing literature dealing with heat stress in Phoenix has established
that physical exposure is linked to reported
health emergencies and human perceptions of
discomfort. Moreover, the social and ecological
structures are intimately intertwined with physical exposure as people with wealth and social
status manipulate their immediate neighborhoods (i.e., through planting trees, maintaining
lawns, and living at lower densities) to reduce
temperatures. Physical exposure thus affects social groups unequally, with socially vulnerable
populations experiencing the most heat. Lacking from these studies, however, is a systematic
spatio-temporal assessment of vulnerability to
heat stress for metropolitan Phoenix, incorporating both biophysical variables and measures
of social sensitivity and resilience.

Study Area and Methodology
Between 1990 and 2000, Phoenix experienced
rapid growth in both population (2.24 to
3.25 million; increase of 45 percent) and land
area (1,223 to 2,069 km2; increase of 69 percent). Changes in surface conditions accompanying rapid urbanization profoundly changed
the local landscape, demography, and ecosystem, with potential consequences for heat vulnerability (Keys, Wentz, and Redman 2007).
Our study specifically focused on the Phoenix
urbanized area as defined by the 2000 U.S. Census. We derived a composite heat vulnerability index (VI total ) from seven measures that are
proxies for physical exposure and social vulnerability or adaptive capacity (Table 1). Given
the complexities of mapping vulnerability with
disparate data sets reviewed in the previous section, it is difficult to determine a priori if any
one measure is of more importance than others. We thus assumed that each measure is of
equal importance in estimating extreme heat
vulnerability.
Near-surface (2 m) maximum (T max ) and
minimum (T min ) air temperatures in June are
indicators of physical exposure. These data
were obtained from thirty-seven meteorological stations distributed across different land-use

types throughout Phoenix. These stations are
part of several well-established, professionally
operated meteorological networks used for
prior UHI analysis (Brazel et al. 2007). June
was selected because average monthly weather
conditions were favorable to UHI development, with numerous hot, clear and calm
days and nights. Further, extreme heat events
occurring in June as opposed to late summer
are more likely to result in heat-related injuries
because residents are less adapted to earlysummer “shocks” (Environmental Protection
Agency [EPA] 2006). Mean June temperatures
were averaged over five years from 1990 to
1994 and 2000 to 2004 to even out the effects
of annual synoptic weather variations. June
maximum temperatures (T max ) usually occur
between 4 and 5 p.m. Conversely, T min, the
minimum temperature, tends to occur around
sunrise, and spatial variations in T min reflect
UHI effects. Higher T max and T min magnitudes
are indicators of greater heat vulnerability.
The normalized differential vegetation index (NDVI) from multispectral satellite imagery has been widely used for measuring
and monitoring plant growth, vegetation cover,
and biomass production (Lillesand, Kiefer, and
Chipman 2004). It utilizes the fact that healthy
vegetation has high (low) reflectance to nearinfrared (visible) bands. NDVI ranges between
±1, with positive values indicating denser surface vegetation. NDVI was derived from Landsat ETM+ images acquired on 19 June 1990
and 14 June 2000 at a resolution of 28.5 m per
pixel, and these represented vegetation density during the summers of 1990 and 2000,
respectively. Areas with high NDVI are assumed to have more effective cooling properties and would reduce exposure and potential
vulnerability.
We selected four measures of social vulnerability to characterize the human system’s capacity to adapt to extreme heat conditions. These
measures were extracted from U.S. Census data
at the census tract level. They reflect an individual’s capacity to adapt to heat as well as the likelihood that the social setting will support the
adaptation process. In terms of individual characteristics, the elderly have been shown to be
more sensitive to heat stress than younger populations because of a biological predisposition
to harm (Smoyer, Rainham, and Hewko 2000).
We thus selected the total population above the

Vulnerability to Extreme Heat in Metropolitan Phoenix 291
Table 1

Measures for vulnerability index to extreme heat, their sources, and what they indicate

Measure
Physical exposure
1. Mean summer maximum
temperature (T max ) from June
1990–1994 and 2000–2004
2. Mean summer minimum
temperature (T min ) from June
1990–1994 and 2000–2004
3. Mean normalized difference
vegetation index (NDVI)
Adaptive capacity
4. Population > 65 years of age
(Pop over65)
5. Median household income
(Med inc)
6. Population of foreign-born
noncitizens (Pop FBNC)
7. Population living in different
residences from 5 years prior
(Diff hous 5)

Source
37 meteorological stations
(Brazel et al. 2007)

Indicator of/relationship to total
vulnerability
Daytime heat stress from regional
climate change/ positive
Nocturnal heat stress from urban heat
island/positive

Landsat ETM+ data

Cooling potential from vegetation
evapotranspiration/negative

U.S. Census

Population most vulnerable to heat
stress/positive
Wealth or poverty/negative

age of sixty-five years (Pop over65) to account
for this tendency to extreme heat vulnerability.
We also used household income (Med inc) as
a surrogate for a household’s ability to use refrigeration or irrigated landscaping to manage
heat stress, as Jenerette et al. (2007) showed that
greater wealth enables Phoenix households to
mitigate heat stress.
The third and fourth indicators measure the
social structure’s ability to mitigate risk when
individuals are physically exposed. The size of
a census tract’s foreign-born, noncitizen population (Pop FBNC) is a proxy for populations
with difficulty in heeding warnings and seeking
help in case of medical emergencies. Harlan et
al. (2006) showed that immigrant status deters
integration within minority neighborhoods.
A notable number of Phoenix’s noncitizen
immigrants are illegal migrants from Mexico
and thus often lack health insurance and access
to the public health care system. Whereas immigrant groups elsewhere have been shown to
have strong internal social networks that foster
social cohesion and community rebuilding
in the face of disaster (e.g., Chamlee-Wright
and Storr 2009; Li et al. 2010), we believe
that the illegal and short-term status of many
of Phoenix’s foreign-born residents weigh in
favor of social isolation, resulting in difficulty
in obtaining resources from established channels of relief. A final vulnerability indicator

Social integration/positive
Social instability/positive

measures the recentness of residence and
population mobility. We assumed that people
who changed residences in the past five years
before the census period (Diff hous 5), such as
short-term renters, were more likely to lack social support in their neighborhoods, hindering
personal access to help during heat-wave events
(EPA 2006). Lack of local surveillance and support combined with social isolation were major
factors in heat deaths in the European heat
wave of 2003 (Fouilett et al. 2006). To summarize, we assumed that high social vulnerability
in a census tract stems from large numbers
of elderly, foreign-born noncitizens, newcomers to the neighborhood, and low-income
residents.
The data were mapped through ESRI
ArcGIS 9.3 mapping software and analyzed
at the census tract scale. Whereas measures 4
through 7 in Table 1 were already formatted
for spatial analysis, raw measures of T max and
T min had to be spatially interpolated through
ordinary kriging, a geostatistical technique
that interpolates the value of a random field for
unobserved locations based on observations at
nearby locations. Subsequently, these kriged
temperature fields were spatially joined to
the census data. We used ENVI 4.6.1, a
software program for processing and analyzing
geospatial imagery, to process layerstaking
and atmosphere correction on raw raster data

292 Volume 64, Number 2, May 2012
Table 2 Correlation coefficient (r) and variance inflation factor (VIF) matrices for raw VItotal measures in
both 1990 (top, n = 458) and 2000 (bottom, n = 646).
Med inc

Pop over65

T min

T max

Diff hous 5

Pop FBNC

NDVI

1.272
–0.149
–0.003
–0.042
–0.040
–0.368
0.251

1.048
–0.079
0.089
0.093
0.017
–0.015

1.114
–0.300
0.060
0.058
–0.127

1.127
–0.072
–0.052
0.139

1.131
0.301
–0.094

1.288
–0.120

1.108

1.581
–0.112
–0.218
–0.110
0.043
–0.476
0.260

1.221
–0.204
0.155
0.260
–0.098
–0.004

1.219
–0.198
–0.187
0.227
–0.035

1.138
–0.026
–0.021
0.172

1.291
0.236
–0.064

1.571
–0.142

1.133

1990
Med inc
Pop over65
Tmin
Tmax
Diff hous 5
Pop FBNC
NDVI
2000
Med inc
Pop over65
Tmin
Tmax
Diff hous 5
Pop FBNC
NDVI

Note: Values in bold are statistically significant at p = 0.01; VIF are italicized. Med inc = median household income;
Pop over65 = population over 65 years old; T min = mean summer minimum temperature; T max = mean summer
maximum temperature; Diff hous 5 = population living in different residences from 5 years prior; Pop FBNC = population
of foreign-born noncitizens; NDVI = normalized difference vegetation index.

from the Landsat ETM+ images. Thereafter,
NDVI was computed by Equation (1):
NDVI =

Band 4 − Band 3
.
Band 4 + Band 3

(1)

where Band 3 is the red band (0.63–0.69 mm)
and Band 4 is the near-infrared (0.63–0.69
mm) of the ETM+ image data. Subsequently,
we used the zonal statistics function in ArcMap
9.3 to obtain both mean temperatures and
NDVI for each census tract.
Once all raw data were geo-referenced and
converted to shapefiles, we examined both correlation coefficient (r) matrices and variance inflation factors (VIF)1 during both periods for
possible multicollinearity among component
variables (Table 2). Six (sixteen) variable pairs
had statistically significant correlations (p =
0.01) in 1990 (2000), although magnitudes of
each bivariate correlation were consistently low
with the exception of Med inc and Pop FBNC
in both periods. The increase in significant correlations between variables over time is most
evident between T min with social vulnerability
indicators, possibly indicating that minor intervariable redundancies exist, particularly with
respect to the social vulnerability component
of VI total (especially in 2000). Although we are
aware of this issue, it should be stressed that

the generally low magnitudes of both r (< 0.5)
and VIF (< 1.581) for all pairs strongly suggest that such impacts were kept to a minimum
and would not largely affect interpretation of
VI total .
Raw vulnerability measures for each tract
were normalized against the entire study area
for both 1990 and 2000 in a ratio treatment used
by Cutter, Mitchell, and Scott (2000) and tested
for normality using the D’Agostino–Pearson
K 2 test. Med inc, however, showed significant
positive skewness after treatment by Equation
(2), and were instead normalized via a ratio of
mean residuals (Equation 3):

xi /
yi =

n



xi

m=1

xmax

yi = zi /zmax ; zi = (xi − x̄) + |xmax | .

(2)
(3)

where yi = normalized vulnerability measure
score in census tract i, xi = raw data of vulnerability measure, x̄ = mean vulnerability measure
for entire study area, xmax = maximum raw vulnerability measure in study area, zi = intermediate vulnerability measure score, and zmax =
maximum intermediate vulnerability score for
entire study area.

Vulnerability to Extreme Heat in Metropolitan Phoenix 293
Normalized measure scores for each tract
varied from 0 to 1, with higher magnitudes
indicating more vulnerability for five of the
component measures (i.e., tracts with higher
raw temperature or greater elderly population
add to total vulnerability). Tracts with higher
raw Med inc (i.e., more wealth) and NDVI
(i.e., more green-space cooling), however, are
negatively related to vulnerability. To account
for this inverse relationship in computing total vulnerability, we used a simple data transform function for normalized data of both
variables (i.e., 1 – yi ). VI total (Equation (4))
was subsequently defined as the linear sum of
all seven normalized vulnerability measures of
equal weight:
V Ito tali = y(Tmax )i + y(Tmin )i
+ (1 − y(NDVI)i)+ (1 − y(Med inc )i )
+ y(Pop over65)i + y(Diff ho us 5)i
+ y(Pop F B NC)i

(4)

Results
Assessing Spatial Distribution of
Vulnerability
In mapping VI total , we divided the metropolitan
area into four quadrants, using the central
business district (CBD) of the City of Phoenix
as the reference point. This is a suitable
landmark as downtown Phoenix remains the
primary economic and transportation node of
the region, which dominates many of the political, social, economic, and cultural aspects of
metropolitan life (Gober and Burns 2002; Keys,
Wentz, and Redman 2007). VI total maps for
1990 and 2000 revealed distinct spatial trends
in vulnerability to extreme heat within Phoenix
(Figure 2). VI total was high in two urban neighborhood types during both periods: (1) the
inner core both in the City of Phoenix and
in several suburban cities in the eastern
metropolitan area (e.g., Tempe, Guadalupe,
Mesa) and (2) within several retirement communities with high concentrations of elderly
(e.g., Sun City, Sun Lakes, and downtown
Mesa). In contrast, several relatively highincome enclaves had persistently low VI total
(e.g., Paradise Valley, Cave Creek, and North
Scottsdale).

Several urban areas underwent significant
changes in VI total between 1990 and 2000 (Figure 3), notably in northwest and southwest
cities, which had increasing VI total trends, especially along the diagonal northwest–southeast
U.S. 60 highway corridor stretching from the
City of Phoenix CBD to the retirement community of Sun City West. Most northeast
and southeast cities, with the exception of
isolated tracts in Scottsdale, South Phoenix,
and South Chandler, experienced decreasing
VI total over the study period, especially in the
fast-growing suburbs of Gilbert and Queen
Creek. A substantial increase of VI total in the
southeast quadrant occurred at Sun Lakes,
a notable elderly enclave where people were
aging into the older than sixty-five years age
cohort.
To explain these patterns, we plotted maps of
normalized scores for individual vulnerability
components to illustrate changing spatial patterns of physical exposure and adaptive capacity
measures in Phoenix (Figures 4 and 5). A strong
desert oasis effect in central Phoenix is seen
for T max in both time periods, with expected
higher T max observed at the urban fringe. UHI
growth is easily discerned with the gradual expansion of higher T min away from the urban
core from 1990 to 2000, with larger than average UHI intensities documented in central
and western Phoenix. The general expansion
of the UHI also corresponded with rapid landcover change resulting from urbanization documented throughout the metropolitan area. Although the central areas of the City of Phoenix
had consistently below-average NDVI, notable
increases in green space were observed in Paradise Valley and North Scottsdale; these are
cities with high-income populations, suggesting possible residential landscape modification
proposed by Jenerette et al. (2007). Conversely,
the gradual conversion of agricultural to suburban land-use, and the resulting decrease in
green space within several southeast and western cities, can also be seen from the 1990 to
2000 NDVI data.
Changing trends in normalized scores for
each adaptive capacity component of VI total
also revealed interesting geographic patterns.
Retirement communities, with many out-ofstate elderly migrants (e.g., Sun City, Sun
Lakes, East Mesa), dot the urban fringe on
the Pop over65 map. Several high-income

294 Volume 64, Number 2, May 2012

Figure 2 (A) 1990 and (B) 2000 VItotal for metropolitan Phoenix. Interval classes are based on standard
deviation of VItotal . (Color figure available online.)

Vulnerability to Extreme Heat in Metropolitan Phoenix 295

Figure 3 
VItotal from 1990 to 2000 for metropolitan Phoenix. Interval classes are based on standard
deviation of VItotal . (Color figure available online.)

northeast cities, such as Paradise Valley and
North Scottsdale, stand out on the Med inc
map in both 1990 and 2000. Several southeast
neighborhoods (e.g., especially in Chandler and
Gilbert) experienced substantial increases in
Med inc as the farming communities, mobile
homes, and retirement villages that once dominated the landscape were replaced with masterplanned communities appealing to more
affluent households. Low-income areas of the
central core persisted and expanded toward the
west from 1990 to 2000.
Sun City, El Mirage, Tempe, central
Phoenix, and Queen Creek were migration foci
for Phoenix’s growing foreign-born, noncitizen population in 1990, and except for
Queen Creek, they were points of settlement
for new immigrants to the region in 2000,
with a notable expansion of predominantly

Mexican-Hispanic neighborhoods in central
and west Phoenix (Gober 2006). Cities at
the urban fringe, especially in the northeast
and northwest, generally had higher concentrations of Diff hous 5 in both periods, reflecting the concentration of new development
there.
VI total declined in the east for a variety of
reasons. First, the sum total of higher income
residents increased; second, it contained a
smaller proportion of foreign-born residents
who concentrated in inner-city neighborhoods
such as within central Phoenix; and third,
it became more stable demographically as
urban growth shifted westward. Residential
areas in the west were also affected more
by larger T max and T min increases relative
to eastern cities, thus increasing physical
exposure.

296 Volume 64, Number 2, May 2012

Figure 4 1990 and 2000 normalized VI maps for physical exposure VItotal components. Tmax = mean
June maximum temperatures; Tmin = mean June minimum temperatures; NDVI = mean June normalized
differential vegetation index. (Color figure available online.)

Vulnerability to Extreme Heat in Metropolitan Phoenix 297

Figure 5 1990 and 2000 normalized VI maps for adaptive capacity VItotal components. Pop over65 =
total population over age 65; Med inc = median household income; Pop FBNC = total population of
foreign born noncitizens; Diff hous 5 = population living in different residences from five years prior.
(Color figure available online.)

298 Volume 64, Number 2, May 2012
Table 3 Vulnerable (defined as +1σ about mean VItotal ) and total population in Phoenix during 1990
and 2000 according to ethnicity

Vulnerable population
1990
2000
Net % change
compared to 1990
baseline
Total population
1990
2000
Net % change
compared to 1990
baseline

African
American/
Black

Hispanic/
Latino

Indian/Alaska
Native

Asian

Non-Hispanic/
Latino White

Total

13,310
19,703
+48.0

69,687
176,471
+153.2

8,097
10,308
+21.5

6,695
8,682
+29.7

285,022
227,838
–20.1

382,811
443,002
+15.7

73,777
113,925
+54.4

337,540
758,176
+124.6

37,570
52,412
+39.5

35,113
66,269
+88.7

1,627,350
2,008,475
+23.4

2,111,350
2,999,257
+42.1

Demographic Variations in Vulnerability
We subsequently examined what these changed
vulnerability maps mean for the region’s racial
and ethnic minorities. We specifically targeted
tracts with VI total one standard deviation above
the metropolitan mean (mean VI total + 1σ ).
This threshold level of high vulnerability was
3.6 in both 1990 and 2000 (Table 3). The sum
of vulnerable tracts was also similar in number
during both time periods (sixty-one in 1990 vs.
sixty-two in 2000). The total metropolitan area
population living in vulnerable areas rose by
16 percent from 1990 to 2000, with the ethnic Hispanic population having the largest rate
of growth at 153 percent, followed by African
Americans with a 48 percent increase. Although
the number of vulnerable people increased for
all minority groups between 1990 and 2000,
the number of vulnerable non-Hispanic Whites
declined from ∼285,000 to ∼228,000, a 20 percent decrease.
These increases in vulnerable populations
must also be compared to corresponding
increases in total population within the study
area to examine proportional demographic
changes over both periods. The proportion of
vulnerable to total population for the entire
metropolitan area decreased from 18.1 to
14.8 percent (Figure 6); non-Hispanic Whites
experienced the largest proportional decrease
of 6.2 percent, followed by a 6 percent decrease
among ethnic Asians, a 1.9 percent drop
among African Americans, and a 0.7 percent
reduction among Native Indian/Alaskan
ethnicities. In contrast, the proportion of

Hispanics living above the VI total threshold of
3.6 increased from 20.6 to 23.3 percent of the
total population, reinforcing the interpretation
that Hispanics were increasingly vulnerable to
heat stress as result of the complex interplay of
physical and social factors.

Discussion
The maps of VI total and its changes between
1990 and 2000 illustrate how climate, urban
ecology, social status, and changing demography interact to create and change the spatial
and temporal patterns of vulnerability to heat
stress. The largest increases of VI total occurred
in western cities within the metropolitan area
where the growth in low-income, foreign-born
populations coincided with increased physical
exposure. High vulnerability increased also in
elderly urban-fringe retirement communities
that were enveloped by the expanding UHI.
Declining vulnerability occurred across eastern
metropolitan cities where population growth
stabilized and UHI development was relatively
weak. These evolving vulnerability surfaces
disproportionately affected the region’s Hispanic population because of the strong indirect
relationships between the four social vulnerability indicators with race and ethnicity. This
was combined with the residential segregation
of Hispanics into neighborhoods with increasing exposure. For instance, historic patterns of
population growth for different ethnic groups
in Phoenix show that Hispanics (mostly Mexican American) concentrated around barrios

Vulnerability to Extreme Heat in Metropolitan Phoenix 299

Figure 6 Proportion of vulnerable ethnic population to total population in 1990 and 2000 within
metropolitan Phoenix (Color figure available online.)

(generally lower class, lower income neighborhoods with a high proportion of Spanishspeaking residents) close to Sky Harbor airport
in the City of Phoenix. To accommodate the
rapid growth of Hispanics, these barrios spread
outward and encompassed much of central
and west Phoenix (Gober 2006). These racial
and ethnic disparities would have been exacerbated if not for non-Hispanic elderly whites
continuing to migrate to Phoenix and settling
into age-segregated retirement communities,
where physical exposure to nighttime heat
was growing from rapid urbanization of the
metropolitan fringe. These communities are
ethnically homogenous areas where more
than 98 percent of residents are non-Hispanic
whites (McHugh and Larson-Keagy 2005).
Despite the influx of this population segment
into retirement neighborhoods, far fewer total
non-Hispanic whites in 2000 were exposed to
vulnerability at above-average thresholds for
heat stress when compared to other ethnic
minorities. Younger, non-Hispanic whites
predominantly settled in several northeast

and southeast cities within the metropolitan
area. These cities experienced declining vulnerability due to (1) low exposure to physical
stresses of high maximum and minimum
temperatures, (2) more neighborhoods with
higher income residents, (3) increased surface
green space through landscape modification
(e.g., in Paradise Valley), and (4) more social
stability from longer term residents. These
physical and social characteristics validate previous, smaller scale findings from Harlan et al.
(2006), who first noted a correlation between
physical exposure and social vulnerability to
heat stress; Jenerette et al. (2007), who linked
socioeconomic status and urban vegetation
with heat stress levels; and Ruddell et al. (2009),
who discerned strong social inequalities with
respect to heat exposure in Phoenix.
Finding that Hispanics in Phoenix are both
socially and physically disadvantaged is consistent with previous research on toxic hazards,
which have also shown the unequal burden carried by minorities who are generally located
adjacent to hazardous places (Bolin et al. 2000;

300 Volume 64, Number 2, May 2012
Bolin et al. 2002; Bolin, Grineski, and Collins
2005; Grineski, Bolin, and Boone 2007). These
studies argue that the burden of minorities is a
function of institutional racism; we do not draw
causal interpretations but note how profoundly
the social and geographic characteristics
of minorities interact to heighten urban vulnerability. These interactions are particularly
worrisome under circumstances of potential
climate change, where the thresholds of extreme heat and resulting harm to health in
the American Southwest will be approached,
if not exceeded more frequently. Vulnerabilities in the physical system could lead to vulnerabilities in the social system, as increasing
exposure to extreme heat would require development of stronger social networks and support systems, which might not keep pace with
changes in magnitudes of physical exposure.

Conclusion
This study assessed the vulnerability of Phoenix
residents in 1990 and 2000 to extreme heat
based on a composite index of vulnerability
(VI total ) based on normalized indexes of physical exposure to heat and several socioeconomic
adaptive capacity measures of equal weight.
We demonstrated that vulnerability varied
significantly over space and time and that it
is unequal across different demographic segments in Phoenix, with Hispanic populations
having a disproportionate exposure to extreme
heat versus other ethnic groups. This marked
disparity is especially apparent in both the
increasing total Hispanic population and the
ratio of Hispanics to total Phoenix population
that were residing in more vulnerable areas
from 1990 to 2000. In contrast, the proportion
of non-Hispanic whites exposed to extreme
heat decreased, despite a large increase in total
elderly migrants in urban-fringe retirement
communities with high VI total .
The need for climate adaptation in Phoenix
is particularly acute, as our research suggests
that many Hispanics in inner-city neighborhoods, and some elderly in retirement
communities, live at loci of heat vulnerability.
The study’s results thus have several policy
implications. First, given the disproportionate
vulnerability to extreme temperatures, policymakers and emergency responders based in

cities or neighborhoods with a large proportion
of vulnerable populations should anticipate
increased HRD calls during heat wave events,
and tailor effective measures for them (e.g.,
more Spanish-speaking responders or specialized elderly medical aid centers). Second,
by identifying areas with high vulnerability,
city officials and policymakers could design
more effective urban adaptation strategies,
such as policies to improve social cohesion
and integration within neighborhoods via
widespread dissemination of heat-stress mitigation information in different languages. 


Note
1

VIF is a measure of the impact of collinearity among
variables in a regression model. VIF > 10 indicates
definite problems of multicollinearity; VIF > 2.5
indicates potential areas of concern. As VIF magnitudes in Table 2 are between 1.0 and 1.6, this
suggests that the collinearity problem among our
independent variables is relatively minor.

Literature Cited
ArcGIS, Version 9.3. Redlands, CA: ESRI, Inc.
Baker, L. C., A. J. Brazel, N. Selover, C. Martin, N.
McIntyre, F. R. Steiner, A. Nelson, and L. Mussacchio. 2002. Urbanization and warming of Phoenix
(Arizona, USA): Impacts, feedbacks, and mitigation. Urban Ecosystems 6:183–203.
Balling, R. C., and S. W. Brazel. 1987. Time and
space characteristics of the Phoenix urban heat island. Journal of the Arizona-Nevada Academy of Science 21:75–81.
Balling, R. C., and R. S. Cerveny. 1987. Long-term
associations between wind speeds and the urban
heat island of Phoenix, Arizona. Journal of Climate
and Applied Meteorology 26:712–16.
Bolin, B., S. Grineski, and T. Collins. 2005.
Geography of despair: Environmental racism
and the making of South Phoenix, Arizona,
USA. Human Ecology Review 12 (2): 155–
67.
Bolin, B., E. Matranga, E. J. Hackett, E. K. Sadalla,
K. D. Pijawka, D. Brewer, and D. Sicotte. 2000.
Environmental equity in a Sunbelt city: The spatial
distribution of toxic hazards in Phoenix, Arizona.
Environmental Hazards 2:11–24.
Bolin, B., A. Nelson, E. Hackett, D. Pijawka, S.
Smith, S. Sadalla, D. Sicotte, and M. O’Donnell.
2002. The ecology of technological risk in a Sunbelt city. Environment and Planning 34:317–39.
Brazel, A. J., P. Gober, S.-J. Lee, S. GrossmanClarke, J. Zehnder, B. Hedquist, and E. Comparri.

Vulnerability to Extreme Heat in Metropolitan Phoenix 301
2007. Determinants of changes in the regional urban heat island in metropolitan Phoenix (Arizona,
USA) between 1990 and 2004. Climate Research
33:171–82.
Brazel, A. J., N. Selover, R. Vose, and G.
Heisler. 2000. Tale of two climates—Baltimore
and Phoenix urban LTER sites. Climate Research
15:123–35.
Centers for Disease Control and Prevention. 2005.
Heat-related mortality: Arizona, 1993–2002, and
the United States, 1979–2002. Morbidity and Mortality Weekly Report 54 (25): 628–30.
Chamlee-Wright, E., and V. H. Storr. 2009. Club
goods and post-disaster community return. Rationality and Society 21 (4): 429–58.
Cutter, S. L., and C. Finch 2008. Temporal and spatial changes in social vulnerability to natural hazards. Proceedings of the National Academy of Sciences
105 (7): 2301–06.
Cutter, S. L., J. T. Mitchell, and M. S. Scott. 2000.
Revealing the vulnerability of people and places: A
case study of Georgetown County, South Carolina.
Annals of the Association of American Geographers 90
(4): 713–37.
Emmanuel, R., and H. J. S. Fernando. 2007. Urban
heat islands in humid and arid climates: Role of
urban form and thermal properties in Colombo,
Sri Lanka and Phoenix, USA. Climate Research
34:241–51.
ENVI, Version 4.6.1. Boulder, CO: ITT Visual Information Solutions.
Environmental Protection Agency (EPA). 2006.
Excessive heat events guidebook. Washington, DC:
U.S. Environmental Protection Agency. http://
www.epa.gov/heatisland/about/heatresponseprogr
ams.html (last accessed 30 June 2009).
Fouilett, A., G. Rey, F. Laurent, G. Pavillon, S. Bellec, C. Guihenneuc-Jouyaux, J. Clavel, E. Jougla,
and D. Hémon. 2006. Excess mortality related to
the August 2003 heat wave in France. International
Archives of Occupational and Environmental Health
80 (1): 16–24.
Gober, P. 2006. Metropolitan Phoenix: Place making
and community building in the desert. Philadelphia:
University of Pennsylvania Press.
Gober, P., A. J. Brazel, R. Quay, S. Myint, S.
Grossman-Clarke, A. Miller, and S. Rossi. 2010.
Using watered landscapes to manipulate urban heat
island effects: How much water will it take to cool
Phoenix? Journal of the American Planning Association 76 (1): 101–21.
Gober, P., and E. K. Burns. 2002. The size and shape
of Phoenix’s urban fringe. Journal of Planning Education and Research 21 (4): 379–90.
Golden, J. S., D. A. Hartz, A. J. Brazel, G. Luber, and
P. Phelan. 2008. A biometeorology study of climate and heat-related morbidity in Phoenix from

2001 to 2006. International Journal of Biometeorology
52:471–80.
Grineski, S., B. Bolin, and C. Boone. 2007. Criteria pollution and marginal populations: Environmental inequity in metropolitan Phoenix, Arizona,
USA. Social Science Quarterly 88 (2): 535–54.
Grossman-Clarke, S., Y. Liu, J. A. Zehnder, and
J. D. Fast. 2008. Simulations of the urban planetary boundary layer in an arid metropolitan
area. Journal of Applied Meteorology and Climatology
47:752–68.
Grossman-Clarke, S., J. A. Zehnder, W. L. Stefanov,
Y. B. Liu, and M. A. Zoldak. 2005. Urban modifications in a mesoscale meteorological model and
the effects on near-surface variables in an arid
metropolitan region. Journal of Applied Meteorology
44 (9): 1281–97.
Guhathakurta, S., and P. Gober. 2007. The impact of
the Phoenix urban heat island on residential water
use. Journal of the American Planning Association 73
(3): 317–29.
Harlan, S. L., A. J. Brazel, L. Prashad, W. L. Stefanov, and L. Larsen. 2006. Neighborhood microclimates and vulnerability to heat stress. Social
Science & Medicine 63:2847–63.
Hawkins, T. W., A. Brazel, W. Stefanov, W. Bigler,
and E. M. Saffell. 2004. The role of rural variability
in urban heat island determination for Phoenix,
Arizona. Journal of Applied Meteorology 43:476–86.
Hedquist, B. C., and A. J. Brazel. 2006. Urban, residential, and rural climate comparisons from mobile transects and fixed stations: Phoenix, Arizona.
Journal of the Arizona-Nevada Academy of Science
38:77–87.
Heinz Center. 2002. Human links to coastal disasters.
Washington, DC: The H. John Heinz III Center
for Science, Economics, and the Environment.
Hsu, S. I. 1984. Variation of an urban heat island in Phoenix. Professional Geographer 36:196–
200.
Jenerette, G. D., S. L. Harlan, A. Brazel, N. Jones,
L. Larsen, and W. L. Stefanov. 2007. Regional relationships between surface temperature, vegetation, and human settlement in a rapidly urbanizing
ecosystem. Landscape Ecology 22:353–65.
Keys, E., E. A. Wentz, and C. L. Redman. 2007.
The spatial structure of land use from 1970–2000
in the Phoenix, Arizona, metropolitan area. The
Professional Geographer 59 (1): 131–47.
Klinenberg, E. 1999. Denaturalizing disaster: A social
autopsy of the 1995 Chicago heat wave. Theory and
Society 28 (2): 239–95.
Li, W., C. A. Airriess, A. C.-C. Chen, K. J. Leong,
and V. Keith. 2010. Katrina and migration: Evacuation and return by African Americans and Vietnamese Americans in an eastern New Orleans suburb. The Professional Geographer 62 (1): 103–18.

302 Volume 64, Number 2, May 2012
Lillesand, T. M., R. W. Kiefer, and J. W. Chipman.
2004. Remote sensing and image interpretation. New
York: Wiley.
Mayhorn, C. B. 2005. Cognitive aging and the processing of hazard information and disaster warnings. Natural Hazards Review 6:165–70.
McHugh, K., and E. M. Larson-Keagy. 2005. These
white walls: The dialectic of retirement communities. Journal of Aging Studies 19 (2): 241–56.
National Research Council. 2006. Facing hazards and
disasters: Understanding human dimensions. Washington, DC: National Academies Press.
Ngo, E. B. 2001. When disasters and age collide: Reviewing vulnerability of the elderly. Natural Hazards Review 2 (2): 80–89.
Oke, T. R. 1982. The energetic basis of the urban
heat island. Quarterly Journal of the Royal Meteorological Society 108:1–24.
Ruddell, D., S. L. Harlan, S. Grossman-Clarke, and
A. Buyantuyev. 2009. Risk and exposure to extreme
heat in microclimates of Phoenix, AZ. In Geospatial techniques to urban hazard and disaster analysis,
ed. P. Showalter and Y. Lu, 179–202. New York:
Springer.
Sailor, D. J. 1995. Simulated urban climate response
to modifications in surface albedo and vegetative cover. Journal of Applied Meteorology 34 (7):
1694–1704.
Semenza, J. C., C. H. Rubin, K. H. Falter, J. D.
Selanikio, W. D. Flanders, H. L. Howe, and J.
L. Wilhelm. 1996. Heat-related deaths during the
July 1995 heat wave in Chicago. New England Journal of Medicine 335 (2): 84–90.
Sheridan, S. C., and L. S. Kalkstein. 2004. Progress
in heat watch-warning system technology. Bulletin
of the American Meteorological Society 85:1931–41.
Smoyer, K. E., D. G. C. Rainham, and J. N. Hewko.
2000. Heat-stress-related mortality in five cities in
southern Ontario: 1980–1996. International Journal
of Biometeorology 44:190–97.
Stabler, L. B., C. A. Martin, and A. J. Brazel. 2005.
Microclimates in a desert city were related to land
use and vegetation index. Urban Forestry & Urban
Greening 3:137–47.
Stone, B., and J. M. Norman. 2006. Land use
planning and surface heat island formation: A
parcel-based radiation flux approach. Atmospheric
Environment 40:3561–73.

Stott, P. A., D. A. Stone, and M. R. Allen. 2004.
Human contribution to the European heatwave of
2003. Nature 432:610–14.
Sun, C.-Y., A. J. Brazel, W. T. L. Chow, B. C.
Hedquist, and L. Prashad. 2009. Desert heat island
study in winter by mobile transect and remote sensing techniques. Theoretical and Applied Climatology
98 (3–4): 323–35.
Turner, B. L., II, P. A. Matson, J. J. McCarthy,
R. W. Corell, L. Christensen, N. Eckley, G. K.
Hovelsrud-Broda, J. X. Kasperson, R. E. Kasperson, and A. Luers. 2003. A framework for vulnerability analysis in sustainability science. Proceedings
of the ‘ Academy of Sciences 100:8080–85.
U.S. Census Bureau. 2007. 50 fastest growing metro areas concentrated in the West and South. Washington,
DC: U.S. Department of Commerce. http://www.
census.gov/Press-Release/www/releases/archives/
population/009865.html (last accessed 30 June
2009).
Wisner, B., P. Blaikie, T. Cannon, and I. Davis.
2004. At risk. 2nd ed. London and New York:
Routledge.

WINSTON T. L. CHOW is a Research Fellow
in the Department of Engineering, Arizona State
University, Mesa, AZ 85212, as well as in the
Department of Geography, National University of
Singapore, Kent Ridge, Singapore 117570. E-mail:
wtchow@asu.edu. His research interests include urban heat islands and sustainable urban climatology.
WEN-CHING CHUANG is a PhD student
in the School of Sustainability, Arizona State
University, Tempe, AZ 85287. E-mail: wenching.chuang@asu.edu. Her research interests include sustainable urban development on linkages of
climate change, urban morphology, and human vulnerability to heat.
PATRICIA GOBER is a Professor of Geography and
Sustainability, as well as a co-director at the Decision
Center for a Desert City, Global Institute of Sustainability, Arizona State University, Tempe, AZ 85287.
E-mail: gober@asu.edu. Her current research focuses
on issues of water management and environmental
change in metropolitan Phoenix.