JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 116, D24113, doi:10.1029/2011JD016720, 2011

An alternative explanation of the semiarid urban area
“oasis effect”
M. Georgescu,1 M. Moustaoui,1 A. Mahalov,1 and J. Dudhia2
Received 12 August 2011; revised 10 October 2011; accepted 12 October 2011; published 22 December 2011.

[1] This research evaluates the climatic summertime representation of the diurnal cycle
of near-surface temperature using the Weather Research and Forecasting System (WRF)
over the rapidly urbanizing and water-vulnerable Phoenix metropolitan area. A suite
of monthly, high-resolution (2 km grid spacing) simulations are conducted during the
month of July with both a contemporary landscape and a hypothetical presettlement
scenario. WRF demonstrates excellent agreement in the representation of the daily to
monthly diurnal cycle of near-surface temperatures, including the accurate simulation
of maximum daytime temperature timing. Thermal sensitivity to anthropogenic land
use and land cover change (LULCC), assessed via replacement of the modern-day
landscape with natural shrubland, is small on the regional scale. The WRF-simulated
characterization of the diurnal cycle, supported by previous observational analyses,
illustrates two distinct and opposing impacts on the urbanized diurnal cycle of the Phoenix
metro area, with evening and nighttime warming partially offset by daytime cooling.
The simulated nighttime urban heat island (UHI) over this semiarid urban complex is
explained by well-known mechanisms (slow release of heat from within the urban fabric
stored during daytime and increased emission of longwave radiation from the urban canopy
toward the surface). During daylight hours, the limited vegetation and dry semidesert
region surrounding metro Phoenix warms at greater rates than the urban complex.
Although prior work has suggested that daytime temperatures are lower within the urban
complex owing to the addition of residential and agricultural irrigation (i.e., “oasis effect”)
we show that modification of Phoenix’s surrounding environment to a biome more
representative of temperate regions eliminates the daytime urban cooling. Our results
indicate that surrounding environmental conditions, including land cover and availability
of soil moisture, play a principal role in establishing the nature and evolution of the diurnal
cycle of near-surface temperature for the greater Phoenix, Arizona, metropolitan area
relative to its rural and undeveloped counterpart.
Citation: Georgescu, M., M. Moustaoui, A. Mahalov, and J. Dudhia (2011), An alternative explanation of the semiarid urban
area “oasis effect”, J. Geophys. Res., 116, D24113, doi:10.1029/2011JD016720.

1. Introduction
[2] Anthropogenic land use and land cover change
(LULCC) can have an impact on the overlying climate via
alteration of the radiation budget and modification of the surface energy balance. A change in surface albedo can alter the
amount of radiative energy absorption while vegetation abundance and water availability (e.g., regions of relatively high
soil moisture) assist in the redistribution of incoming energy

1
School of Mathematical and Statistical Sciences, Center for
Environmental Fluid Dynamics, Global Institute of Sustainability, Arizona
State University, Tempe, Arizona, USA.
2
Mesoscale and Microscale Meteorology Division, National Center for
Atmospheric Research, Boulder, Colorado, USA.

Copyright 2011 by the American Geophysical Union.
0148-0227/11/2011JD016720

into latent and sensible heating back into the atmosphere
[Pielke, 2001]. Urbanization is an extreme manifestation of
anthropogenic LULCC that can further modulate the nearsurface climate [Oke, 1987], and by means of a phenomenon
known as the urban heat island (UHI) effect has been shown
to increase urban relative to rural near-surface temperatures [Bornstein, 1968; Oke, 1973; Landsberg, 1981]. Largescale urbanization, resulting from changes in associated
surface properties, can therefore influence local-to-regionalscale climates by modifying the dynamic and thermodynamic properties of a region.
[3] The UHI effect has been the subject of research for
many years [Arnfield, 2003; Brazel et al., 2000; Gedzelman
et al., 2003; Hawkins et al., 2004; Oke and Maxwell, 1975;
Sailor, 1995; Taha et al., 1991; Taha, 1997; Tarleton and
Katz, 1995; Tran et al., 2006; Zhou et al., 2004]. Generally regarded as an evening and nighttime phenomenon
[Oke, 1987], the UHI has also been shown to influence

D24113

1 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

daytime temperatures [e.g., Tran et al., 2006] and, in some
cases, to have a daytime signature that exceeds that of its
nighttime counterpart [e.g., Cheval and Dumitrescu, 2009].
[4] Urban canyon geometry plays a key role in regulating
both the daytime and nighttime UHIs [Oke, 1981]. Because
of the reduction of unobstructed visible sky within an urban
canopy, longwave radiation loss is reduced after sunset,
leading to greater near-surface temperatures relative to rural
environs during evening and nighttime hours. The geometric
configuration of urban areas can have the opposite effect
during daytime, as incoming radiation is reduced because of
shading effects [Oke, 1987]. The greater heat capacity of the
built environment can further intensify the nighttime UHI as
stored energy is slowly released during the night [Grimmond
and Oke, 1999].
[5] During daylight hours, the presence of urban parks can
decrease local temperatures [Spronken-Smith et al., 2000] by
means of repartitioning toward greater latent as opposed to
sensible heating, leading to an oasis effect [Oke, 1987]. The
oasis effect (i.e., repartitioning toward greater latent heating
because of the presence of irrigated lawns, parks, etc.) has
been proposed as a likely mechanism for lower urban daytime temperatures over semiarid cities (e.g., Phoenix, Arizona)
compared with the surrounding semidesert environment
[Brazel et al., 2000; Golden et al., 2006; Stabler et al., 2005],
although debate as to the existence and extent of this mechanism continues [Fast et al., 2005]. It is possible that the oasis
effect depends on individual city characteristics, including
urban morphology, building design, and structured materials
used, as well as time of season and particular climatic
regimes [Potchter et al., 2006]. That the oasis effect has a
local impact is not in question [Spronken-Smith et al., 2000],
but whether local cooling by means of this mechanism
scales beyond the microscale is an important subject that
warrants attention. Research addressing the existence and
mechanism of the semiarid urban area daytime cool island
phenomenon is essential for land use and city planners,
energy producers, and inhabitants of semiarid cities, requiring improved understanding of the principal consequences
of continued LULCC and associated urbanization.
[6] Urban complexes located in semiarid regions, such as
the geater Phoenix, Arizona, metropolitan area, are particularly vulnerable as UHI-related impacts aggravate the already
severe nature of semiarid environments. Water sustainability
concerns resulting from rising demands that are due to population increase [Gober and Kirkwood, 2010] and projected
impacts from large-scale anthropogenic climate change on
large watersheds such as the Colorado River Basin [e.g.,
Barnett and Pierce, 2008; Rajagopalan et al., 2009] requires
a fundamental understanding of impacts (e.g., effects on the
diurnal cycle of near-surface temperature) resulting from
continued urbanization in these already-stressed environments.
[7] The Phoenix metropolitan area, located in the semiarid Sonoran desert of the southwestern United States, has
witnessed extensive land surface modification since the
middle of the previous century [Knowles-Yanez et al., 1999;
Georgescu et al., 2009], with considerable impacts on nearsurface temperatures [Brazel et al., 2000]. Analysis of historical temperature records indicates that Phoenix exemplifies among the most extreme instances of urban-induced
warming [Hansen et al., 1999]. Of particular concern is the
region’s ever-increasing urban heat island (UHI) intensity

D24113

[Hsu, 1984; Balling and Brazel, 1987; Brazel et al., 2000],
resulting from unrelenting development and urbanization.
Consequences for UHI enhancement extend beyond the climate realm, raising air quality [Balling et al., 2001; Jacobson,
2010], ecological [Grimm and Redman, 2004], social [Harlan
et al., 2008], and sustainability [Gober and Kirkwood, 2010]
concerns. Further growth into a megapolitan complex known
as the Sun Corridor is projected to bring the region’s population total to near 10 million by 2040 [Gammage et al., 2008].
Monitored and sustainable expansion is therefore required
[Grimm et al., 2008], and impact assessment of historical
urban development and growth is a key step toward ensuring
minimum climate-ecosystem impacts of the region’s relentlessly expanding urban footprint.
[8] The principal aim of the current study is to evaluate the
ability of the Weather Research and Forecasting System
(WRF), applied over the semiarid Phoenix metropolitan
area, to accurately represent the evolution of the diurnal
cycle of near-surface temperatures. A key question relates to
the principal mechanism(s) responsible for the urban area’s
daytime cool island. By focusing on an individual urban
hub, this work is likely to shed further light on precise localto-regional-scale consequences associated with urban expansion. Our objective is to build upon previous work, focusing
on improving our understanding of the development of
the diurnal cycle of near-surface temperatures [e.g., Zehnder,
2002; Grossman-Clarke et al., 2005; Georgescu et al., 2008,
2009] to better characterize impacts of urbanization and better
illustrate the sensitivity of local-to-regional-scale climate to
LULCC over this semiarid metropolitan complex. We focus
on the summertime climatic evolution of the Phoenix diurnal
cycle of near-surface temperatures, which is of significant
importance to millions of the region’s inhabitants, as stress
levels (indicated by combined effects of extreme temperatures
and moisture levels) are at their maximum during this time
of year [Golden et al., 2008].

2. Methods
2.1. RCM Model Used: WRF-ARW
[9] We have used version 3.2.1 of the WRF [Skamarock
et al., 2008; Skamarock and Klemp, 2008] modeling system to conduct high-resolution (2 km grid spacing within
the innermost domain) sensitivity experiments to anthropogenic LULCC over the semiarid Phoenix metropolitan area.
The ability of WRF to reproduce the region’s midsummer
climate was evaluated against appropriate observations (see
section 2.2). A detailed description and inventory of options
utilized, including domain extent, forcing data, parameterization options, and irrigation application are presented below.
[10] WRF is a fully compressible, nonhydrostatic model
that has been used for a variety of applications, ranging from
local to global scales. We initialized and forced WRF, at
the lateral boundaries, with data obtained from the North
American Regional Reanalysis (NARR) data set [Mesinger
et al., 2006]. Simulations for all experiments were initialized on 30 June, 00:00 UT, of the respective year and concluded on 31 July, 12:00 UT, of the respective year. We used
three nested grids, centered over the Phoenix Sky Harbor
International Airport (Figure 1a).
[11] The four-layer Noah land surface model (LSM)
[Chen and Dudhia, 2001] was used as the land surface

2 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

D24113

component of WRF to update soil temperature and moisture
following model initialization. The Noah LSM scheme has
been widely used in the regional climate modeling community as an operational scheme at the National Centers for
Environmental Prediction (NCEP), in the development of the
25 year NARR data set [Mesinger et al., 2006], and as part of
an international program focused on producing future regional
climate change assessments (North American Regional Climate Change Assessment Program: http://narccap.ucar.edu/
data/rcm-characteristics.html). To better account for urbanrelated processes, we made use of the single-layer Noah Urban
Canopy Model (UCM) [Kusaka and Kimura, 2004], which
has been evaluated in both coupled [Lo et al., 2007] and
uncoupled modes [Loridan et al., 2010].
[12] The U.S. Geological Survey’s (USGS) 21 class, 30 m
1992 National Land Cover Data set (NLCD92) [Vogelmann
et al., 2001] was used to represent modern-day LULC within
the Noah LSM. The landscape classification scheme used
is presented in Table 1.
[13] Irrigated agriculture was accounted for by setting the
volumetric water content of cropland category pixels (class
12 in Table 1; see also Figure 1b), for all soil layers, to soil
saturation as a function of the green vegetation fraction (i.e.,
water is not added to bare soil for which vegetation is
lacking). Notwithstanding the semiarid nature of the region,
the major crops of Maricopa County (home to the Phoenix
metro area) include cotton (growing season ranges from
April to October), vegetables (e.g., soybean; growing season
ranges from June through late fall), and orchards used for
oranges and grapefruits (annual growing season). In each
case, significant irrigation is required for crop growth. For
example, at full blossom peak (i.e., during midsummer),
cotton requires up to 25 mm day 1 of irrigation water [Erie
et al., 1982]. Irrigation was set at a frequency of once daily
for cropland category pixels. A simple drip-style irrigation
scheme was implemented for the low-intensity residential
class within WRF to account for the region’s common
method of mesic landscaping. Every third day, the second
(10–40 cm depth) and third (40–100 cm depth) soil layers of
the Noah LSM corresponding to this urban class were set
to a reference volumetric soil moisture (i.e., threshold below
which transpiration begins to stress but above wilting point)
level, also as a function of the green vegetation fraction. The
green vegetation fraction for each grid cell was based on a
monthly, 5 year, 0.144°, climatological data set assumed valid
on the 15th of the month [Gutman and Ignatov, 1998] and
interpolated to daily values. Albedo was accounted for by
means of inclusion of a monthly, 5 year, 0.144°, climatological
data set assumed valid on the 15th of the month, also interpolated to daily values [Csiszar and Gutman, 1999].

Figure 1. (a) Domain extent and geographical representation of the WRF nested grid configuration with topography
(in meters) overlaid. (b) NLCD92 landscape representation of
grid 3 used in WRF simulations. Individual station locations
used to evaluate WRF are marked: E, Encanto; G, Greenway;
S, Sky Harbor International Airport; L, Litchfield; El, Eloy;
M, Maricopa. (c) Pre-Settlement landscape representation of
grid 3 used in WRF simulations.

3 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

Table 1. LULC Categories Used for Noah LSM
LULC

Noah

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
31
32
33

Evergreen Needleleaf
Evergreen Broadleaf
Deciduous Needleleaf
Deciduous Broadleaf
Mixed Forests
Closed Shrublands
Open Shrublands
Woody Savannas
Savannas
Grasslands
Permanent Wetland
Croplands
Cropland or Natural Vegetation Mosaic
Snow and Ice
Barren or Sparsely Vegetated
Ocean
Wooded Tundra
Mixed Tundra
Bare Ground Tundra
Urban Low-Intensity Residential
Urban High-Intensity Residential
Urban Commercial or Industrial

2.2. Experiments Performed
[14] We used a nested grid configuration with all three
grids centered over the Phoenix Sky Harbor International
Airport (Figure 1a). The finest grid (grid 3) utilized a spacing of 2 km in the horizontal direction, with increasing
mesh size for the intermediate (grid 2; 8 km) and coarsest
grid (grid 1; 32 km) (Figure 1b). It is important to note that
our intent is not to capture effects of individual urban elements within the built environment, but to resolve the landatmosphere interactions and associated dynamics relevant
to the local-to-regional scale LULCC for the metropolitan
region. We made use of a modified version of the KainFritsch convective parameterization scheme to represent
convective processes on the coarsest grid only [Kain, 2004]
and represented convective motions explicitly on grids 2
and 3.
[15] Guided by previous research, we restricted our WRF
sensitivity (to LULCC) experiments to separate dry monsoon
seasons, as the impact of the land surface may more readily
be distinguished during such regimes [e.g., Anyah et al.,
2008; Georgescu et al., 2008, 2009]. In fashioning our
experimental design this way, we simulated three different
Julys, corresponding to three different years (thus accounting for interannual variability) using two different land cover
representations (Table 2). Monsoon season selection was
based on NARR data availability (1979 to present day) and
analysis of the Daily Unified Precipitation data set [Higgins
et al., 2000]. July months corresponding to 1979, 1989, and
1994 were chosen (Table 2). The precise methods and selection criteria used are discussed in detail by Georgescu [2008].
[16] Modern-day urban areas were represented via use of
the 1992 USGS-derived 30 m NLCD92 data set [Vogelmann
et al., 2001] (see section 2.1 for details). Produced as part
of Federal Region 9 (which includes the states of Arizona,
California, and Nevada), the Arizona portion of the NLCD92
data set was based on satellite imagery obtained between the
late 1980s and early 1990s. The region’s LULC representation
was therefore based on this same time, and we selected the

D24113

month-year of model evaluation (July 1990) to be consistent
with the timing of underlying land surface representation.
[17] A hypothetical land surface was developed by replacing all anthropogenic pixels (this consisted of three different categories of urban land use and irrigated agriculture; see
also Table 1) with native vegetation (i.e., shrubland; hereafter presettlement) (Figure 1c). Monthly mean differences
between the three (July) simulations with the NLCD92 and
presettlement landscape scenarios (referred to as ensemble
differences for the purpose of this study) enabled direct
comparison of midsummer climate sensitivity to anthropogenic landscape change.
2.3. Observational Data
[18] The source of near-surface temperature data against
which WRF was evaluated is described below.
[19] We make use of five regional stations available from
the Arizona Meteorological Network (AZMET: http://ag.
arizona.edu/azmet) and a sixth station, Phoenix Sky Harbor
International Airport (the only first-order National Weather
Service station in the region), courtesy of the National Climatic Data Center. Of the six stations, three are urban
stations (Phoenix Encanto, Phoenix Greenway, and Sky
Harbor International Airport) while the remaining three
are located outside the metropolitan area (Eloy, Litchfield,
and Maricopa). The locations of each station relative to one
another and the boundary walls of the innermost grid are
shown in Figure 1b. It is important to note that although our
experiments are high resolution in nature (2 km on the
innermost grid), one cannot expect exceptional agreement
between a model-resolved 4 km2 grid cell and a station
location covering an area of 1 m2. For this reason, observations are averaged over all (six) stations and compared with
the corresponding average of model-simulated, nearest grid
point station locations, from 1 July 1990 (12:00 UT) through
31 July 1990 (12:00 UT). We discard the initial 36 simulation hours and utilize this time period as model spin-up.
Observational data availability with a temporal frequency
of 3 h and equivalent model output generation allows for
assessment of the diurnal cycle.

3. Results
3.1. Evaluation of WRF
[20] We begin evaluation of WRF’s performance via
comparison against near-surface temperature observations
(see section 2.3 above). WRF performs reasonably during
the course of the month, tracking the day-to-day variability
of near-surface temperatures with notable fidelity (Figure 2).
The generally well-simulated daytime warming is balanced
Table 2. Summary of Experiments Performeda
Landscape Representation

WRFb

NLCD92
NLCD92
Presettlement

1990c
1979, 1989, 1994
1979, 1989, 1994

a
For each experiment, the analysis time consists of the period lasting
from 1 July, 12:00 UT, through 31 July, 12:00 UT.
b
Years from which initial and lateral boundary conditions were used to
force WRF in June/July.
c
Denotes experiment used as control simulation and evaluated against
observations.

4 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

D24113

interacting elements of the land-atmosphere system essential
for achieving the correct amplitude and phase of these forced
modes. In particular, an accurate representation of the evolving nature of the diurnal cycle serves as a key examination of
many aspects of the model’s physical parameterizations.
Although it is beyond the scope of this paper, sensitivity of
the simulated diurnal cycle over this semiarid region to the
choice of physical parameterizations is an essential question
that is the scope of future research.

Figure 2. Observed (black contour) and WRF-simulated
(red contour) 2 m air temperature for 1 July 1990 (12:00 UT)–
31 July 1990 (12:00 UT).
by equally well-simulated nighttime cooling, resulting in a
diurnal range that is of similar magnitude to observations.
This difference is apparent when quantifying impacts on
urban stations only (Table 3) (we assess impacts on urban
stations only as it is these stations that are most influenced
by utility of the UCM scheme). Although WRF does display
a small warm bias over urban locales, the representation of
the monthly averaged diurnal range of near- surface temperatures is well captured.
[21] The simulation of amplitudes and phases of the diurnal
and monthly cycles of near-surface temperatures provides
an ideal test bed for model parameterizations and representation of the interacting surface, boundary layer, and free
atmosphere. To more fully examine WRF performance, the
near-surface temperature profiles presented in Figure 2 are
filtered to separate diurnal variations and longer period fluctuations. We apply a band-pass nonrecursive filter using a
Kaizer window to avoid the Gibbs phenomenon [Hamming,
1983], spanning periods ranging between 18 and 30 h. The
diurnal variability simulated by WRF closely follows the
observed variability (Figure 3a). WRF is able to properly
resolve both the timing of the diurnal cycle (Figure 3b) and
the temperature fluctuations after applying a low-pass filter
that removes all periods less than 30 h (Figure 3c).
[22] Figure 3b presents the composite, monthly averaged
diurnal cycle for diurnal variations of temperature derived
for both WRF and observations. These average cycles are
calculated by selecting an hour during a day and averaging
the temperature perturbation values found at that hour
between different days of the month. The composite cycle
illustrates the correct WRF-simulated evolution of the diurnal cycle, in a mean sense, including the timing of peak
daytime temperatures. The standard deviation (Figure 3d)
associated with the WRF-simulated cycle, with respect to
observations, highlights the relative range of near-surface
temperature uncertainty, depicting an envelope of model
variability, relative to observations, within 1°C (Figure 3d).
[23] The relative performance of WRF on both diurnal and
monthly time scales reflects the proper representation among

3.2. Sensitivity to LULCC
[24] Guided by the well-represented diurnal cycle presented
in the previous section, we next assess the impact of nearsurface temperature sensitivity to LULCC. Near-surface temperature sensitivity to anthropogenic LULCC closely follows
the pattern of landscape modification, with areas undergoing
urbanization experiencing diurnally averaged warming that is
partly offset by conversion to irrigated agriculture (Figure 4a).
Warming and cooling impacts of LULCC largely offset
each other in a regional sense. To account for the influence
of LULCC on the monthly averaged diurnal cycle, we next
present distinct land use conversion themes (e.g., examination
of grid cells undergoing conversion from shrub (presettlement
scenario) to urban (NLCD92) land) to evaluate how each
cover’s alteration has contributed to changes in the evolution
of the diurnal cycle (Figure 4b). The simulated conversion of
native shrubland to irrigated agriculture imparts a peak (daytime) cooling of 0.15°C. This impact (on the near-surface
temperature) appears to be of secondary importance to that of
urban development. The conversion of native shrubland to
urban land drives two distinct and opposing impacts during
the course of the diurnal cycle: cooling during the daytime
hours and warming during the evening and nighttime hours
(i.e., the well-known UHI effect). Mean evening and nighttime warming (peaking at 2.6°C) is partially offset by mean
daytime cooling (peaking at 1.3°C) and results in an
overall warming approaching 1°C for grid cells undergoing
conversion to urban land.
[25] The accurate representation of the diurnal cycle
and its agreement with observations (Figures 2 and 3 and
section 3.1) provides confidence that the simulated daytime
cooling (which emerges on the climatic timescale (Figure 4b))
resulting from urbanization is a robust feature, rather than a
model artifact.
[26] Averaged daily variations of near-surface temperature
differences (NLCD92-presettlement) for grid cells undergoing landscape conversion from shrubland to urban land
are presented for midafternoon and late-afternoon time

Table 3. Mean Observed and WRF-Simulated Maximum,
Minimum, and Difference Between Maximum and Minimum 2 m
Air Temperature Across the Trio of Urban Stationsa

Tmean
THi
TLo
THi TLo

Observations

WRF

32.9
38.3
27.6
10.7

33.7 (+0.8)
39.6 (+1.3)
28.3 (+0.7)
11.3 (+0.6)

a
The number in parentheses denotes departure relative to observations.
Analysis is for the period 1 July 1990 (12:00 UT) through 31 July 1990
(12:00 UT).

5 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

D24113

Figure 3. (a) Two meter air temperature perturbation, after applying a band-pass filter spanning the range
between 18 and 30 h, for observations (black contour) and WRF (red contour). (b) Composite monthly
averaged diurnal cycle of band-pass filtered 2 m air temperature for observations (black contour) and
WRF (red contour). (c) As in Figure 3a, but after applying low-pass filter that removes all periods less than
30 h. (d) Simulated 2 m air temperature 1 standard deviation of the composite monthly averaged diurnal
cycle (red contours) relative to observations (black contour).
periods to illustrate the aforementioned daytime cooling and
transition to urban warming that occurs during late afternoon and early evening (Figure 5a). Differences between
NLCD92 and presettlement are greater for midafternoon
hours (21:00 UT), with presettlement temperatures consistently warmer, by about 1°C, during the course of July. By
late afternoon (00:00 UT), these differences are reduced by
half, fluctuating between 0°C and 1°C during the course of
the month (warmer for presettlement relative to NLCD92).
Further support for this conclusion is available from previous observational analyses by Brazel et al. [2000], who
document historical, negative, (May) urban minus rural
near-surface maximum (i.e., daytime) temperature differences between a number of urban and residential stations
and a small, rural town located 25 km southeast of the

metropolitan area that has remained mostly undeveloped
[Brazel et al., 2000, Figure 4c]. Although there are methodological differences in the (modeling) approach used in
this paper and the approach used by Brazel et al. [2000]
(entirely observational), their results explicitly illustrate the
daytime urban area cooling relative to outlying desert locations (i.e., undeveloped land) that was simulated in our
experiments. Figure 5b presents an analysis similar to that of
Figure 5a, but for nighttime hours. Near-surface temperature
differences between NLCD92 and presettlement are generally between 1.5°C and 3°C, highlighting the greater simulated
warming impact of the nighttime UHI relative to daytime
cooling . As before, additional observational support for this
conclusion is available from prior analyses of observational
data documenting positive nighttime temperature differences

6 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

D24113

between the NLCD92 ensemble members and presettlement
ensemble members. The pattern of landscape modification
is evident, with the largest differences in sensible heating
occurring over areas where land was modified from shrubland to irrigated agriculture and urban land (also compare
with Figures 1b and 1c to note areas of anthropogenic
LULCC). LULCC resulted in an overall decrease in sensible
heating. For areas undergoing a conversion from semidesert
to irrigated agriculture, the decrease in sensible heating was
compensated by an increase in latent heating of similar
magnitude (Figure 6b). The presence of soil water and irrigation, required to sustain agricultural productivity in this
harsh environment, resulted in a repartitioning of energy
toward enhanced latent heating. Differences in latent heat flux,
however, were negligible over areas undergoing urbanization,

Figure 4. (a) WRF-simulated ensemble-averaged 2 m air
temperature difference (°C) (monthly mean of NLCD92
members minus monthly mean of presettlement members).
(b) WRF-simulated time series of diurnally averaged 2 m
air temperature difference (°C) for grid cells undergoing
conversion from shrub (presettlement scenarios) to urban
(NLCD92 scenarios) (dashed red line), shrub (presettlement
scenarios) to irrigated agriculture (NLCD92 scenarios)
(dashed blue line), and all pixels (presettlement scenarios)
to all pixels (NLCD92 scenarios) (solid black line).
between urban and surrounding desert landscapes [Brazel
et al., 2000, Figure 4d].
3.3. Impact of LULCC on Energy Balance
[27] We begin the discussion of the causes underlying
impacts on the diurnal cycle of near-surface temperatures by
examination of the energy balance components. Figure 6a
presents monthly averaged sensible heat flux differences

Figure 5. (a) WRF-simulated 2 m temperature difference
for grid cells undergoing conversion from shrub (presettlement scenarios) to urban (NLCD92 scenarios) land during
daytime, at 21:00 UT and 00:00 UT. (b) As in Figure 5a,
but for nighttime at 09:00 UT (blue line) and 12:00 UT
(red line).

7 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

D24113

and normalizing against net all-wave radiation did not alter
this conclusion (not shown). Turning off irrigation within the
low-intensity residential area did not indicate any discernible
influence on near-surface temperatures or on the components
of the energy balance (not shown).
[28] In addition to repartitioning surface-absorbed energy
between latent and sensible heating, the relatively high heat
capacity of the urban fabric is responsible for storing a significant fraction of available energy [Grimmond and Oke,
1999]. Differences in monthly averaged ground heat flux
between the NLCD92 and presettlement simulations are
shown in Figure 6c. Negative differences indicate greater heat
storage for NLCD92 relative to the presettlement simulations
(ground heat flux is positive when directed away from the
surface and negative when directed into the surface; greater
negative values of ground heat flux are simulated for
NLCD92 simulations relative to the presettlement experiments, resulting in the negative differences presented in
Figure 6c). As the Sun rises each morning, rather than heating
surface and near-surface air as would happen over the nondeveloped and sparsely vegetated surrounding semidesert,
more of the incoming energy is lost through ground heat flux
resulting from the physical mass of the built environment.
[29] The increase in urban heat storage (relative to the
undeveloped semidesert), which occurs largely during daylight hours, results in increased sensible heat fluxes during
the nighttime hours for NLCD92 relative to presettlement
(Figure 7a). The slow release of stored energy is a principal
factor for the urban-induced nighttime warming noted in the
preceding section (see Figure 4b). The warmer nighttime
urban canopy is also responsible for the increasing emission
of longwave energy toward the surface (resulting from a
reduction in the sky view factor, leading to a decrease in net
longwave radiation loss to the atmosphere), although this
effect appears to be largest over the industrial-commercial
landscape and diminishes rapidly over lower-intensity residential areas (Figure 7b).
3.4. Daytime Cool Island
[30] The previous section discussed impacts of LULCC
and associated urbanization on the surface energy balance
but did not address possible reasons for the simulated daytime cooling effect besides heat storage within the urban
infrastructure. That the built environment results in enhanced
heat storage is not a distinguishing feature of semiarid cities
relative to their temperate brethren [Grimmond and Oke,
1999]. Therefore, we next address the basis for the simulated daytime cooling resulting from urbanization.
[31] Figure 8 shows monthly averaged, WRF-simulated 2
m air temperatures across latitude band 33.5°N for the paired
(i.e., NLCD92 and presettlement) simulations, for each of
the trio of simulated Julys, at 11:00 local standard time
(LST) (that time of day when the simulated cooling was
greatest; see also Figure 4b). The particular latitude band

Figure 6. (a) Monthly averaged WRF-simulated sensible
heat flux difference (mean of NLCD92 members minus
mean of presettlement members). (b) As in Figure 6a, but
for latent heat flux. (c) As in Figure 6a, but for ground heat
flux. Units are W m 2.
8 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

D24113

(along the selected transect, irrigated agriculture begins just
east of about 112.5°W while the developed urban area starts
at about 112.3°W; see also Figure 1b). The sharp decrease in
near-surface temperatures for the paired simulations just east

Figure 7. (a) As in Figure 6a, except for nighttime hours.
(b) Nighttime simulated downward longwave flux difference
at surface (mean of NLCD92 members minus mean of presettlement members). Nighttime hours: 20:00 LST–05:00 LST.
chosen crossed through the surrounding semidesert and traversed through the urban core. Diagnosing contrasting
semidesert and urban impacts was deemed necessary, and
deciding on latitude bands slightly north or south of 33.5°N
did not modify the interpretation of our results.
[32] It is evident, from an assessment of Figure 8, that
near-surface temperatures between the paired simulations
are similar west of the anthropogenically modified landscape

Figure 8. Monthly averaged WRF-simulated cross section
of 2 m air temperature at 33.5°N at 11:00 LST for NLCD92
scenarios (solid blue line) and presettlement scenarios (solid
red line) for (a) 1979, (b) 1989, and (c) 1994. The y axis on
right-hand side represents varying terrain heights (solid
black line) across a given latitude band.

9 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

Figure 9. As Figure 8 but for NLCD92 scenarios (solid
blue line) and NLCD92FOREST scenarios (solid red line) for
(a) 1979, (b) 1989, and (c) 1994.
of 112.6°W is related to changes in topography; this is also
the case east of about 111.7°W. Differences in near-surface
temperatures among the paired simulations (for all three
simulated Julys) begin as one proceeds toward the irrigated
agriculture region, with warming for the sparsely vegetated

D24113

desert (i.e., presettlement scenario) relative to irrigated agriculture (i.e., the NLCD92 scenario) of a couple of tenths of
a degree Celsius. Continued progression eastward (e.g., to
about 112.2°W, where one is firmly entrenched within the
urban area) results in a further and relatively larger decrease
in temperature (compared with the progression from semidesert land to irrigated agriculture land). While surfaceabsorbed energy was efficiently used to heat the near-surface
air over the vegetation-sparse and water-limited semidesert
simulated in the presettlement scenarios, there is significantly
less energy used to heat air over the urban region, in large
part because of the enhanced heat storage of the built environment (see section 3.3). The region where urbanization
occurred also coincided with the lowest cross-sectional elevation, resulting in the largest near-surface temperature
contrasts between the presettlement and NLCD92 simulations. These differences are consistent among the differently
simulated Julys, and the molar-shaped temperature pattern
for each of the modern-day landscapes (i.e., NLCD92 for
each of the trio of Julys) is lower relative to the surrounding semidesert (e.g., west of 112.4°W and east of about
111.9°W, notwithstanding temperature variability that is due
to topography).
[33] We posit that the decrease in daytime temperature
over the urban area (relative to outlying areas) noted upon
urbanization (i.e., the simulated cooling) occurs because of
greater heating rates in the surrounding semidesert and is
related to the biome in which the city resides and not to any
unique features of metro Phoenix. To test this hypothesis,
we conducted three additional simulations using the modernday representation of greater Phoenix with the following
changes. We altered the surrounding semidesert biome to a
temperate broadleaf forest biome that is more distinctive of
major urban areas in, for example, the northeastern United
States. To do this, the shrubland category surrounding metro
Phoenix was replaced with an evergreen broadleaf landscape. The sparse vegetation (generally at 20% vegetation
fraction or less) of the semidesert is not fitting for a temperate broadleaf forest biome and consequently required
modification. The vegetation fraction was set to a constant
value of 90% for the broadleaf forest and irrigated agriculture categories (this agrees well with July forest coverage
west of, for example, Boston, Massachusetts, and with July
vegetation coverage of agricultural areas in the Midwest
United States) [Gutman and Ignatov, 1998]. Last, the initial
soil moisture value was increased by 75% for all soil layers
to more accurately represent the nondesiccated soils of a
temperate broadleaf landscape. All other settings remained
identical to the trio of NLCD92 simulations, including irrigation frequency within the urban area and over irrigated
agriculture locations.
[34] Figure 9, presented in a fashion similar to that of
Figure 8, shows results of the new NLCD92 experiments
with metro Phoenix located within a temperate broadleaf
forest biome (hereafter referred to as NLCD92FOREST)
alongside the previously documented NLCD92 simulations,
with metro Phoenix located within its natural semidesert
environment. The general trend of lower near-surface air
temperatures is apparent for NLCD92FOREST compared with
NLCD92. This result is not surprising given the repartitioning toward greater latent relative to sensible heating.

10 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

Notably, the urban area is no longer cooler than its surrounding environment. Rather, for each of the trio of simulated Julys, metro Phoenix is as warm or warmer than its
neighboring surroundings, as the rapid daytime heating over
the parched semidesert adjoining Phoenix has been replaced
with a landscape that warms gradually, in a manner more
appropriate for temperate regions.

4. Discussion and Concluding Remarks
[35] We have performed an evaluation of and sensitivity
to anthropogenic LULCC using the WRF regional climate
model, applied over the semiarid and water-vulnerable
Phoenix metropolitan area, with emphasis on the depiction
of the evolving diurnal cycle of near-surface temperatures
and representation of the urban area’s daytime cool island.
[36] Our results demonstrate an excellent depiction of the
daily and monthly evolution of the midsummer diurnal cycle
with WRF. Low-level thermal day-to-day variability,
monthly averaged composite diurnal cycles, and standard
deviations (relative to observations) illustrate the correct
representation of the semiarid region’s harsh summer climate. The proper simulation of peak daytime temperature
timing is of particular importance to hot and arid urban areas
such as Phoenix that require accurate and timely profiles of
peak energy demand associated with basic human adaptation
and function.
[37] Although near-surface temperature sensitivity to
anthropogenic LULCC is small on the regional scale, our
experiments, supported by previous observational analyses,
show two distinct and opposing impacts on the urbanized
diurnal cycle during the extent of the full day, with evening
and nighttime warming (relative to the undeveloped semidesert landscape) partially offset by daytime cooling. To our
knowledge, these are the first simulations to explicitly depict
the daytime cooling effect of the urbanizing metro Phoenix
landscape.
[38] Analysis of the surface energy budget components
indicates a shift toward greater latent heating for regions
undergoing a change to irrigated agriculture and enhanced
heat storage for regions undergoing urbanization. The slow
release of stored daytime heat and enhanced emission of
longwave radiation toward the surface (from within the
urban canopy) for urbanizing areas during the evening and
nighttime hours is largely responsible for the nighttime
warming influence. This result is in agreement with previous
modeling [e.g., Grossman-Clarke et al., 2005] and observationally based research [e.g., Brazel et al., 2000]. Contrary to
prior speculation, however, our results do not indicate a significant impact on near-surface temperatures (and surface
energy budget components) because of irrigation within urban
areas, suggesting that the oasis effect is not the principal
mechanism explaining the urban area’s daytime cooling.
[39] To determine the cause of the observed and simulated
cooling, the semidesert biome within which greater Phoenix
resides was replaced with a temperate broadleaf forest
biome. Modification of metro Phoenix’s surrounding environment eliminated the urban cooling entirely, suggesting
that adjacent conditions (e.g., soil moisture, land cover) play
a principal role in establishing the nature and evolution of
the diurnal cycle of near-surface temperatures for the metro

D24113

area. This result is in agreement with recent observationally
based evidence showing the dependence of the UHI of
diverse cities across the United States on differing ecological
settings and environmental factors [Imhoff et al., 2010]. The
precise impact of urbanization on the diurnal cycle of nearsurface temperatures largely depends on the surrounding
environment in which urbanization occurs, including landscape and soil moisture conditions, rather than any distinguishing features of the city itself.
[40] Future work quantifying the impact of urbanization on
the diurnal cycle of near-surface temperatures should focus
on other rapidly expanding, and water-vulnerable, semiarid
metropolitan regions (e.g., Las Vegas, Nevada, or Riyadh,
Saudi Arabia). It is plausible that urbanization impacts on the
diurnal cycle illustrate similar daytime cooling and nighttime
warming (of greater relative magnitude) as documented over
metro Phoenix, potentially distinguishing general urbanization effects over semiarid metropolitan regions from their
counterparts in temperate zones.
[41] Given further anticipated urban expansion and LULCC
over many regions, attention should also be given to sustainably built urban forms [Jo et al., 2010; Seto et al., 2010]
that mitigate the severity of urban-induced warming. This
is important for many currently developing metropolitan
regions, but is especially critical for water-vulnerable metro
areas such as Phoenix. These urban hubs are currently facing significant water sustainability concerns that are due to
the rising demands associated with rapid population increase
[Gober and Kirkwood, 2010], with stringent urban adaptation
measures potentially resulting in future allocation reductions.
[42] Acknowledgments. This work was funded by NSF grant ATM0934592.

References
Anyah, R. O., C. P. Weaver, G. Miguez-Macho, Y. Fan, and A. Robock (2008),
Incorporating water table dynamics in climate modeling: 3. Simulated
groundwater influence on coupled land-atmosphere variability, J. Geophys.
Res., 113, D07103, doi:10.1029/2007JD009087.
Arnfield, A. J. (2003), Two decades of urban climate research: A review
of turbulence, exchanges of energy and water, and the urban heat island,
Int. J. Climatol., 23, 1–26, doi:10.1002/joc.859.
Balling, R. C., Jr., and S. W. Brazel (1987), Time and space characteristics
of the Phoenix urban heat island, J. Ariz. Nev. Acad. Sci., 21, 75–81.
Balling, R. C., Jr., R. S. Cerveny, and C. D. Idso (2001), Does the urban
CO2 dome of Phoenix, Arizona contribute to its heat island?, Geophys.
Res. Lett., 28, 4599–4601, doi:10.1029/2000GL012632.
Barnett, T. P., and D. W. Pierce (2008), When will Lake Mead go dry?,
Water Resour. Res., 44, W03201, doi:10.1029/2007WR006704.
Bornstein, R. D. (1968), Observations of the urban heat island effect in New
York City, J. Appl. Meteorol., 7, 575–582, doi:10.1175/1520-0450(1968)
007<0575:OOTUHI>2.0.CO;2.
Brazel, A., N. Selover, R. Vose, and G. Heisler (2000), The tale of two
climates–Baltimore and Phoenix urban LTER sites, Clim. Res., 15,
123–135, doi:10.3354/cr015123.
Chen, F., and J. Dudhia (2001), Coupling an advanced land-surface/hydrology
model with the Penn State/NCAR MM5 modeling system. Part I: Model
description and implementation, Mon. Weather Rev., 129, 569–585,
doi:10.1175/1520-0493(2001)129<0569:CAALSH>2.0.CO;2.
Cheval, S., and A. Dumitrescu (2009), The July urban heat island of
Bucharest as derived from MODIS images, Theor. Appl. Climatol., 96,
145–153, doi:10.1007/s00704-008-0019-3.
Csiszar, I., and G. Gutman (1999), Mapping global land surface albedo
from NOAA AVHRR, J. Geophys. Res., 104, 6215–6228, doi:10.1029/
1998JD200090.
Erie, L. J., O. F. French, D. A. Bucks, and K. Harris (1982), Consumptive
use of water by major crops in the Southwestern United States, Agric.

11 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

Res. Ser. Conserv. Res. Rep. 29, 40 pp., U.S. Dep. of Agric., Washington,
D. C.
Fast, J. D., J. C. Torcolini, and R. Redman (2005), Pseudovertical temperature profiles and the urban heat island measured by a temperature datalogger network in Phoenix, Arizona, J. Appl. Meteorol., 44, 3–13,
doi:10.1175/JAM-2176.1.
Gammage, G., Jr., J. S. Hall, R. E. Lang, R. Melnick, and N. Welch (2008),
Megapolitan: Arizona’s Sun Corridor, Morrison Institute for Public
Policy, Ariz. State Univ., Tempe.
Gedzelman, S. D., S. Austin, R. Cermak, N. Stefano, S. Partridge,
S. Quesenberry, and D. A. Robinson (2003), Mesoscale aspects of the urban
heat island around New York City, Theor. Appl. Climatol., 75, 29–42.
Georgescu, M. (2008), Evaluating the effect of land-use and land-cover
change on climate in the greater Phoenix, AZ, region, Ph.D. dissertation,
Ariz. State Univ., Tempe.
Georgescu, M., G. Miguez-Macho, L. T. Steyaert, and C. P. Weaver (2008),
Sensitivity of summer climate to anthropogenic land-cover change over
the Greater Phoenix, AZ, region, J. Arid Environ., 72, 1358–1373,
doi:10.1016/j.jaridenv.2008.01.004.
Georgescu, M., G. Miguez-Macho, L. T. Steyaert, and C. P. Weaver (2009),
Climatic effects of 30 years of landscape change over the Greater Phoenix,
AZ, region: 1. Surface energy budget changes, J. Geophys. Res., 114,
D05110, doi:10.1029/2008JD010745.
Gober, P. A., and C. W. Kirkwood (2010), Vulnerability assessment of climateinduced water shortage in Phoenix, Proc. Natl. Acad. Sci. U. S. A., 107,
21,295–21,299, doi:10.1073/pnas.0911113107.
Golden, J., A. Brazel, J. Salmond, and D. Laws (2006), Energy and water
sustainability: The role of urban climate change from metropolitan infrastructure, J. Eng. Sustain. Dev., 1, 55–70.
Golden, J. S., D. Hartz, A. Brazel, G. Luber, and P. Phelan (2008), A biometeorology study of climate and heat-related morbidity in Phoenix from
2001 to 2006, Int. J. Biometeorol., 52, 471–480, doi:10.1007/s00484007-0142-3.
Grimm, N. B., and C. L. Redman (2004), Approaches to the study of urban
ecosystems: The case of central Arizona-Phoenix, Urban Ecosyst., 7,
199–213, doi:10.1023/B:UECO.0000044036.59953.a1.
Grimm, N. B., D. Foster, P. Groffman, J. M. Grove, C. S. Hopkinson,
K. J. Nadelhoffer, D. E. Pataki, and D. P. C. Peters (2008), The changing
landscape: Ecosystem responses to urbanization and pollution across
climatic and societal gradients, Front. Ecol. Environ, 6(5), 264–272,
doi:10.1890/070147.
Grimmond, C. S. B., and T. R. Oke (1999), Heat storage in urban
areas: Local-scale observations and evaluation of a simple model, J. Appl.
Meteorol., 38, 922–940, doi:10.1175/1520-0450(1999)038<0922:HSIUAL>
2.0.CO;2.
Grossman-Clarke, S., J. A. Zehnder, W. L. Stefanov, Y. 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, J. Appl. Meteorol., 44, 1281–1297, doi:10.1175/JAM2286.1.
Gutman, G., and A. Ignatov (1998), The derivation of green vegetation fraction
from NOAA/AVHRR data for use in numerical weather prediction models,
Int. J. Remote Sens., 19, 1533–1543, doi:10.1080/014311698215333.
Hamming, R. W. (1983), Digital Filters, Prentice-Hall, Englewood Cliffs,
N. J.
Hansen, J., R. Ruedy, J. Glascoe, and M. Sato (1999), GISS analysis of
surface temperature change, J. Geophys. Res., 104(D24), 30,997–31,022,
doi:10.1029/1999JD900835.
Harlan, S. L., A. Brazel, G. D. Jenerette, N. S. Jones, L. Larsen, L. Prashad,
and W. L. Stefanov (2008), In the shade of affluence: The inequitable
distribution of the urban heat island, Res. Soc. Probl. Public Policy, 15,
173–202, doi:10.1016/S0196-1152(07)15005-5.
Hawkins, T. W., A. J. Brazel, W. L. Stefanov, W. Bigler, and E. M. Saffell
(2004), The role of rural variability in urban heat island determination for
Phoenix, Arizona, J. Appl. Meteorol., 43, 476–486, doi:10.1175/15200450(2004)043<0476:TRORVI>2.0.CO;2.
Higgins, R. W., et al. (2000), Improved US precipitation quality control system and analysis, NCEP/Clim. Predict. Cent. ATLAS 7, 40 pp., Natl.
Cent. for Environ. Predict., Camp Springs, Md.
Hsu, S. I. (1984), Variation of an urban heat island in Phoenix, Prof.
Geogr., 36, 196–200, doi:10.1111/j.0033-0124.1984.00196.x.
Imhoff, M. L., P. Zhang, R. E. Wolfe, and L. Bounoua (2010), Remote sensing of the urban heat island effect across biomes in the continental USA,
Remote Sens. Environ., 114, 504–513, doi:10.1016/j.rse.2009.10.008.
Jacobson, M. Z. (2010), Enhancement of local air pollution by urban air
domes, Environ. Sci. Technol., 44, 2497–2502, doi:10.1021/es903018m.
Jo, J. H., J. Carlson, J. S. Golden, and H. Bryan (2010), Sustainable urban
energy: Development of a mesoscale assessment model for solar reflective
roof technologies, Energ. Policy, 38, 7951–7959, doi:10.1016/j.enpol.2010.
09.016.

D24113

Kain, J. S. (2004), The Kain–Fritsch convective parameterization: An update,
J. Appl. Meteorol., 43, 170–181, doi:10.1175/1520-0450(2004)043<0170:
TKCPAU>2.0.CO;2.
Knowles-Yanez, K., C. Moritz, J. Fry, C. L. Redman, M. Bucchin, and
P. H. McCartney (1999), Historic Land Use Team: Phase I Report on
Generalized Land Use, Cent. Ariz.–Phoenix Long-Term Ecol. Res.
Contrib. 1, 21 pp., Arizona State Univ., Tempe.
Kusaka, H., and F. Kimura (2004), Thermal effects of urban canyon structure on the nocturnal heat island: Numerical experiment using a mesoscale model coupled with an urban canopy model, J. Appl. Meteorol.,
43, 1899–1910, doi:10.1175/JAM2169.1.
Landsberg, H. E. (1981), The Urban Climate, 278 pp., Academic, San
Diego, Calif.
Lo, J. C. F., A. K. H. Lau, F. Chen, J. C. H. Fung, and K. K. M. Leung
(2007), Urban modification in a mesoscale model and the effects on
the local circulation in the Pearl River Delta region, J. Appl. Meteorol.
Climatol., 46, 457–476, doi:10.1175/JAM2477.1.
Loridan, T., C. S. B. Grimmond, S. Grossman-Clarke, F. Chen, M. Tewari,
K. Manning, A. Martilli, H. Kusaka, and M. Best (2010), Trade-offs and
responsiveness of the single-layer urban canopy parametrization in WRF:
An offline evaluation using the MOSCEM optimization algorithm and
field observations, Q. J. R. Meteorol. Soc., 136, 997–1019, doi:10.1002/
qj.614.
Mesinger, F., et al. (2006), North American regional reanalysis, Bull. Am.
Meteorol. Soc., 87, 343–360, doi:10.1175/BAMS-87-3-343.
Oke, T. R. (1973), City size and the urban heat island, Atmos. Environ., 7,
769–779, doi:10.1016/0004-6981(73)90140-6.
Oke, T. R. (1981), Canyon geometry and the nocturnal urban heat
island: Comparison of scale model and field observations, J. Climatol.,
1, 237–254, doi:10.1002/joc.3370010304.
Oke, T. R. (1987), Boundary Layer Climates, 2nd ed., 435 pp., Methuen,
London.
Oke, T. R., and G. B. Maxwell (1975), Urban heat island dynamics in
Montreal and Vancouver, Atmos. Environ., 9, 191–200, doi:10.1016/
0004-6981(75)90067-0.
Pielke, R. A. (2001), Influence of the spatial distribution of vegetation
and soils on the prediction of cumulus convective rainfall, Rev. Geophys.,
39, 151–177, doi:10.1029/1999RG000072.
Potchter, O., P. Cohen, and A. Bitan (2006), Climatic behavior of various
urban parks during hot and humid summer in the Mediterranean city of
Tel Aviv, Israel, Int. J. Climatol., 26, 1695–1711, doi:10.1002/joc.1330.
Rajagopalan, B., K. Nowak, J. Prairie, M. Hoerling, B. Harding, J. Barsugli,
A. Ray, and B. Udall (2009), Water supply risk on the Colorado River: Can
management mitigate?, Water Resour. Res., 45, W08201, doi:10.1029/
2008WR007652.
Sailor, D. J. (1995), Simulated urban climate response to modification in
surface albedo and vegetative cover, J. Appl. Meteorol., 34, 1694–1704,
doi:10.1175/1520-0450-34.7.1694.
Seto, K. C., R. Sanchez-Rodriguez, and M. Fragkias (2010), The new geography of contemporary urbanization and the environment, Annu. Rev.
Environ. Resour., 35, 167–194, doi:10.1146/annurev-environ-100809125336.
Skamarock, W. C., and J. B. Klemp (2008), A time-split nonhydrostatic
atmospheric model for research and NWP applications, J. Comput. Phys.,
227, 3465–3485, doi:10.1016/j.jcp.2007.01.037.
Skamarock, W. C., J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, M. G.
Duda, X.-Y. Huang, W. Wang, and J. G. Powers (2008), A description
of the advanced research WRF version 3, NCAR Tech. Note NCAR/TN–
475+STR, 125 pp., Natl. Cent. for Atmos. Res., Boulder, Colo.
Spronken-Smith, R. A., T. R. Oke, and W. P. Lowry (2000), Advection and the surface energy balance across an irrigated urban park,
Int. J. Climatol., 20, 1033–1047, doi:10.1002/1097-0088(200007)20:9<
1033::AID-JOC508>3.0.CO;2-U.
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 For.
Urban Green., 3, 137–147, doi:10.1016/j.ufug.2004.11.001.
Taha, H. (1997), Urban climates and heat islands: Albedo, evapotranspiration, and anthropogenic heat, Energy Build., 25, 99–103, doi:10.1016/
S0378-7788(96)00999-1.
Taha, H., H. Akbari, and A. Rosenfeld (1991), Heat island and oasis effects
of vegetative canopies: Micrometeorological field measurements, Theor.
Appl. Climatol., 44, 123–138, doi:10.1007/BF00867999.
Tarleton, L. F., and R. W. Katz (1995), Statistical explanation for trends in
extreme summer temperatures at Phoenix, Arizona, J. Clim., 8, 1704–1708,
doi:10.1175/1520-0442(1995)008<1704:SEFTIE>2.0.CO;2.
Tran, H., D. Uchihama, S. Ochi, and Y. Yasuoka (2006), Assessment
with satellite data of the urban heat island effects in Asian mega cities,
Int. J. Appl. Earth Obs. Geoinf., 8, 34–48, doi:10.1016/j.jag.2005.05.003.

12 of 13

D24113

GEORGESCU ET AL.: THE SEMIARID URBAN AREA “OASIS EFFECT”

Vogelmann, J. E., S. M. Howard, L. Yang, C. R. Larson, B. K. Wylie, and
J. N. Van Driel (2001), Completion of the 1990s National Land Cover
Data Set for the conterminous United States, Photogramm. Eng. Remote
Sens., 67, 650–662.
Zehnder, J. A. (2002), Simple modifications to improve fifth-generation
Pennsylvania State University-National Center for Atmospheric Research
Mesoscale Model performance for the Phoenix, Arizona, metropolitan area,
J. Appl. Meteorol., 41, 971–979, doi:10.1175/1520-0450(2002)041<0971:
SMTIFG>2.0.CO;2.
Zhou, L., R. E. Dickinson, Y. Tian, J. Fang, Q. Li, R. K. Kaufmann,
C. J. Tucker, and R. B. Myneni (2004), Evidence for a significant

D24113

urbanization effect on climate in China, Proc. Natl. Acad. Sci. U. S. A.,
101(26), 9540–9544, doi:10.1073/pnas.0400357101.
J. Dudhia, Mesoscale and Microscale Meteorology Division, National Center
for Atmospheric Research, PO Box 3000, Boulder, CO 80307–3000, USA.
M. Georgescu, A. Mahalov, and M. Moustaoui, School of Mathematical
and Statistical Sciences, Center for Environmental Fluid Dynamics, Global
Institute of Sustainability, Arizona State University, Tempe, AZ 85287,
USA. (matei.georgescu@asu.edu)

13 of 13