1. Introduction
The dimensions of individual elements such as buildings, streets, sidewalks, courtyards, and urban parkland determine microclimate length scales, which may extend from less than 1 m to hundreds of meters (Oke 2004a). Microclimates show enhanced variability in space and time relative to local-scale (102–104 m) climates (Grimmond et al. 2004). Based on the size of the surface area characterized by the model presented here (1260 m2), a 1-km2 area may include up to 800 distinct microclimates.
Microscale effects dominate surface energy fluxes and climatological variables within the urban canopy layer (UCL), at length scales similar to those of individual structures (Pearlmutter and Berliner 2005; Grimmond et al. 2004). The UCL extends from ground level to mean building height. The roughness sublayer is that part of the atmospheric surface layer where flow is dynamically influenced by individual roughness elements. Typically, flow in the surface layer is smoothed at twice to several times the height of individual roughness elements, which is represented by the height of the UCL in built-up urban areas (Grimmond et al. 2004; Mills 1997a). Measurements taken at an appropriate height above the roughness sublayer will represent local-scale effects, and below this height, microscale effects (Oke 2004b; Offerle et al. 2003).
This study presents a semiempirical energy balance model that is not overly demanding in terms of its instrumentation. It calculates energy fluxes through time at the microclimatic scale and can be used as a tool to investigate microclimate responses to variations in spatial, temporal, or synoptic conditions, or to material properties. It is applicable to a surface area of approximately 1260 m2, based on the height of the radiometer above the surface. This spatial scale is similar to dimensions of urban features such as buildings, streets, and courtyards. Energy fluxes for natural surfaces can also be calculated with this model. It incorporates conventional meteorological and radiation measurements recorded with portable, easy-to-use instruments, and it calculates surface energy fluxes with high statistical significance.
Surface energy flux calculations are useful for a variety of purposes. The model proposed here could calculate surface energy fluxes and temperatures for a variety of rooftop designs, including reflective, green, and conventional. It could also be used to evaluate heat mitigation techniques, such as reflective surfaces or shade trees, by assessing surface temperature responses to individual energy components modified by the techniques (Bretz et al. 1998; Akbari et al. 1999, 2001). Installing instruments in a developing urban area could help to determine desirable forms and locations for built structures based on an understanding of the relation between urban microclimates, architectural design, and urban planning (Evans and DeSchiller 1996; DeSchiller and Evans 1996; Golany 1996). Empirical model results are useful tools for validation of more complex urban climate models (Masson 2006; Mills 1997a; Kanda et al. 2005). Building heating and cooling loads are directly related to material (roof and wall) surface temperatures, which are influenced by energy balance components specifically impacting them (Santamouris et al. 2001; Akbari et al. 1999).
2. Methods
This study characterizes microclimates associated with two urban rooftop sites. Chicago City Hall was one site and the other was a building at The University of Chicago. Figure 1 is a map of the two sites within the city of Chicago. The instruments were placed on rooftops to ensure their safety and for logistical reasons that prohibited placement at street level within the dense urban fabric of downtown Chicago near City Hall. Site and instrument descriptions are presented in this section, as well as certain considerations for conducting intercomparisons between two distinctive urban sites.
a. City Hall—high-rise urban
City Hall is a 12-story (66 m) building of primarily limestone construction. Many buildings within 1 km2 of City Hall are over 20 stories tall, or 2–3 times its height. Lake Michigan is 1.5 km east of the site. Approximately 16% of its ∼5000-m2 rooftop is covered with highly reflective PVC polymer impermeable membrane, 64% is a garden, or green, rooftop, and the remaining 20% opens to courtyards below. It is located at 41°53′N, 87°37′W (land elevation 179 m MSL), in a densely developed urban area characterized by close-set high-rise buildings (Fig. 2). A mirror-image rooftop directly adjoins City Hall, creating a total footprint area of ∼10 000 m2. However, that rooftop is composed of black tar asphalt. The instruments were placed in a safe corner on the reflective coating between a brick wall, the black rooftop, and a courtyard edge (Fig. 3). This site is termed “City Hall.”
b. University of Chicago—low-rise urban
The University building is 5 stories tall (25 m) and consists of brick and limestone construction. The rooftop is black pitch material covered with gray pebbles approximately 3 cm deep. It is located at 41°47′N, 87°36′W (land elevation 179 m MSL), in a highly developed area. The surrounding area is characterized by three- to seven-story buildings of similar construction materials and grassy areas such as parks and courtyards (Fig. 4). A close-up view of this sampling location is shown in Fig. 5. A rooftop edge was located 3.5 m west of the instruments. City Hall is 10.6 km north of the University site, and Lake Michigan is 1.9 km east. This site is termed “University.”
c. Instruments
Identical net radiometers and weather stations collected contemporaneous measurements from the two sites. The Kipp and Zonen Model CNR-1 net radiometer consists of upward- and downward-facing pyranometers, as well as separate pyrgeometers to measure solar shortwave and thermal longwave irradiance (W m−2), respectively, from the sky and underlying surface. The upward-facing sensors sample the entire 180° hemisphere, whereas the downward-facing sensors sample an area proportional to the height of the instrument above the surface. For a height (H) of 2 m (used here), 99% of the input for the downward-facing sensors comes from a circular area with a radius of 10H, or 1256.7 m2. This sampling area defines the length scales under consideration.
The Davis Vantage Pro2 weather station includes a rain collector, air temperature and humidity sensors, an anemometer, and a solar radiation sensor. The temperature and humidity sensors have an accuracy of ±0.5°C and ±3%, respectively. The tipping bucket reads rainfall amounts in 0.2-mm increments. The anemometer measures wind direction and speeds ranging from 0 to 76.1 m s−1. A fan-aspirated radiation shield reduces effects of solar and rooftop heating of the instrument, which might otherwise affect air temperature readings.
To assure data comparability, both sets of instruments were collocated at University for 15 days, recording measurements at 10-min intervals from 21 September to 5 October 2005. Collocation occurred after instruments were removed from City Hall. The University set was installed after emplacement of the City Hall set, so this presented the first opportunity for collocation. Plots of measurements from like instruments are presented in Fig. 6, and associated regression equations and R2 values are listed in Table 1. We concluded that both sets of instruments recorded compatible measurements under identical climate conditions, so differences between recorded variables are due to environmental circumstances and not to instrument nonequivalence.
Physical features within the sampling area radius will impact radiometer and weather measurements. For example, the nearby wall, edge, and adjoining rooftop at City Hall are within the sampling area radius. One consequence of this is that the albedo of the City Hall sampling area is not a function of the white rooftop alone. University also includes wall and edge effects, though to a lesser extent than at City Hall. Structural features are common to urban rooftops and will affect both radiometer and meteorological measurements, but precisely determining the effects of these features, singly and jointly, on surface energy balance components is a complex problem. This paper attempts to identify effects of microscale features on energy fluxes by comparing model results from two sites that experience similar synoptic conditions, but differ in site-specific physical features with dimensions on the order of <1 km.
d. Site intercomparison
3. Energy balance model of urban surfaces
A one-dimensional energy balance model provides a framework for interpreting the measurements described in the previous section. The instantaneous temperature of a surface exposed to the atmosphere is determined by the balance between several energy fluxes that individually act to heat or cool that surface. This section develops the semiempirical mathematical model, which statistically derives energy fluxes that were not measured directly. A comparison of fluxes from different environments, such as the City Hall and University sites, demonstrates the physical mechanisms responsible for observed differences in surface temperature.
Consider a solid slab of thickness L (not related to Lowry’s L) oriented either horizontally, representing the ground or a rooftop, or vertically, representing the vertical face of a building. The slab thickness measured along the x axis extends from x = −L to x = 0, and the slab is infinite and homogeneous along both axes perpendicular to x. The face at x = 0 is exposed to the atmosphere, while the face at x = −L is either underground or constitutes the interior wall of a structure. The temperature of the slab as a function of x over the range x = −L to x = 0 at time t is T(x, t), and the surface temperature of the slab is TS(t) = T(0, t). The domain x > 0 corresponds to the atmosphere, which exchanges energy with the slab. The temperature of the atmosphere in the immediate vicinity of the surface is TA(t) at time t. Figure 7 is a diagram of the model geometry.
Five processes lead to the exchange of energy between the slab and the region x > 0, and these are described by fluxes, expressed in watts per meter squared into or out of the surface at x = 0. The fluxes arise from 1) absorption of shortwave solar energy at the surface of the slab, 2) absorption of incoming thermal longwave energy at the surface of the slab, 3) emission of thermal longwave energy at the surface of the slab, 4) exchange of sensible heat between the slab and the atmosphere, and possibly 5) evaporative cooling, provided that liquid water is present on the surface of the slab. By convention, a flux of energy that flows in the +x direction is positive. This applies to fluxes directed from the surface of the slab into the atmosphere and from the interior of the slab toward the surface at x = 0. A negative flux corresponds to flow from the atmosphere into the slab surface or to flow within the slab toward the interior surface at x = −L.
The incoming solar and thermal radiation fields as well as the shortwave albedo “α” of the slab surface are measured directly by the net radiometers. The absorbed solar and thermal energy fluxes are −(1− α)K(t) and −L(t), respectively, at time t, where the minus signs indicate flow from the atmosphere to the surface. The flux of thermal radiation emitted by the slab surface is εσTS(t) 4, where ε is the emissivity of the material, usually very close to 1.0, and σ = 5.67 × 10−8 W m−2 K−4 is the Stefan–Boltzmann constant.
The model assumes the flux of sensible heat between the solid surface and the atmosphere to be proportional to the temperature contrast and equal to QH(t) = S(t)[TS(t) – TA(t)] (Arnfield and Grimmond 1998; Pearlmutter et al. 1999). The transport coefficient S(t) is a parameter to be estimated from measurements, and in general it will vary in time with wind speed. The magnitude of the energy flux due to evaporative cooling is QE(t), and the simple temperature-dependent form adopted here is QE(t) = h0 + h1TS(t), where h0 and h1 are estimated from measurements. For some urban surfaces, such as the sides of buildings or dry rooftops, QE(t) = 0 since no liquid water is present.
The constant term (h0 − q0) and regression coefficients (h1 − q1, β, S0, and S1) in Eqs. (13) and (14) are estimated via least squares regression applied to radiometer (Q*, TS) and weather (TA, υ) measurements obtained over at least 24 h. This yields estimates of QH and (QE − ΔQS) that, when taken with the measured radiative fluxes, provide a description of the surface energy balance. Because of their similar mathematical forms, terms for QE and ΔQS are factored together such that results yield an estimate for the net effect of these two fluxes (i.e., QE − ΔQS). If QE is assumed negligible because of the absence of liquid water, then the derived value would be based only on ΔQS.
4. Results
This section presents an intercomparison of energy fluxes and temperatures between the two sites and describes how these parameters are related to site-specific features. Three assumptions warrant explanation. First, QE is assumed to be negligible, since no surface liquid water was present and no precipitation occurred. This implies that the derived value of (QE − ΔQS) is based on conduction. The second assumption is that emissivity, ε, is equal to unity. While this is not strictly true, both surfaces are expected to have ε in excess of 0.9. The third assumption is that daytime fluxes determine nocturnal cooling patterns. To capture the daytime-heating, nocturnal-cooling pattern for a full 24 h, results are presented with a start time of 0400 LT, which precedes dawn by approximately 2 h.
For this analysis we used values of Q*(t), TS(t), TA(t), and υ(t) measured at 10-min intervals from 0400 29 July to 0400 5 August 2005 at both University and City Hall. At both sites, TA is measured at 2 m above the underlying surface. Least squares regression applied to Eqs. (13) and (14) estimates the coefficients h0 − q0, h1 − q1, β, S0, and S1. From these we infer time-dependent values for the flux of sensible heat, QH(t) = [S0 + S1υ(t)][TS(t) − TA(t)], and the net effect of evaporative cooling (assumed negligible) and heat conduction, QE(t) − ΔQS(t) = (h0 − q0) + (h1 − q1)TS(t) + β dTS(t)/dt. Table 2 lists the coefficients and percent of variance explained by the regression when applied to the measurements collected from each site. The regression model explains 96.0% of the variance in net radiation observed at University and 83.7% at City Hall. All derived coefficients have high statistical significance, with the ratio of the best estimate to standard error always exceeding 3.
Fluxes from the 7 consecutive days were averaged into a 24-h period to illustrate the diurnal features of the surface energy balance for the 2 sites. Table 3 lists the 24-h average values. Table 4 presents the corresponding average surface and air temperatures and wind speeds. The reported surface temperatures are calculated assuming an emissivity equal to 1 (ε = 1) and by applying the Stefan–Boltzmann radiation law to the L↑ measurements recorded at 10-min intervals. These averages differ slightly from what would arise if the surface emissivities and surface temperature were known (Stewart et al. 1994; Sun and Mahrt 1995). Table 5 lists the mean daily values for TA, K↓, and υ for each of the seven days. Days 1–6 had mostly clear to partly cloudy conditions, and day 7 was mostly cloudy.
No precipitation occurred during the week, and physical inspections of the sampling sites confirmed that surfaces were continuously dry. Therefore, based on the sampling period measurements, QE can be assumed negligible; however, its representation by the model cannot be evaluated in this context. Barzyk (2006) reports that QE is in fact represented in the term QE − ΔQS, supported by model results from a green rooftop, rural grassland, a snow-melting event, a suburban yard, and various wet–dry scenarios that occurred throughout a year-long sampling period.
a. Solar radiation
The largest difference between the two sites is in the values of K↓, as illustrated in Fig. 8a. Table 3 lists the 24-h average values for this component. Rooftop structures and surrounding high-rise buildings obstruct the direct solar beam incident on the City Hall radiometer, leading to a value of K↓ that is approximately 100 W m−2 less than that from University. A City Hall rooftop wall adjacent to the radiometer accounts for approximately half this value, and surrounding high-rise buildings, the other half. Figure 8a demonstrates that sunrise in an intensely developed high-rise urban area may occur later than in less developed areas because of artificial obstructions of K↓; and sunset may occur sooner for the same reason. These two phenomena are termed “urban sunrise” and “urban sunset,” respectively.
The reflective potential of the City Hall white rooftop membrane is effectively decreased because of its proximity to the adjacent black tar rooftop of the adjoining county building, a nearby brick wall, and edge effects. While reflective membranes typically have α > 0.65 when new, and ≥ 0.5 through time, the measured average albedo at City Hall was 0.36. University has an albedo of 0.23, a value that arises primarily from the underlying surface. During periods of direct solar irradiance, when the albedo of the white rooftop was most effective, in this case at ∼1300 LT (Fig. 8b), City Hall reflected an additional ∼120 W m−2 of K↓ relative to University, which demonstrates the cooling potential of this highly reflective rooftop.
b. Thermal radiation
Table 3 and Fig. 9 show that differences in solar heating are partially offset by enhanced L↓ at City Hall, which is ∼30 W m−2 greater than at University throughout the 24-h period. The average atmospheric radiating temperature based on σT 4 = L↓ at City Hall is ∼20°C and at University is ∼15°C. Assuming that the vertical structures around City Hall have values for TS similar to those found at University (i.e., 30.4°C) and that the value of L↓ at University is representative of the unobstructed atmosphere, then L↓ recorded at City Hall would result if approximately two-thirds of it were received from the atmosphere and one-third from built structures. Fisheye photography could be used to substantiate this (Grimmond et al. 2001), but it was unavailable; however, visual inspection of the City Hall sky view suggests that these fractions are reasonable.
c. Conduction
Figure 10 illustrates the values of this energy component throughout the 24-h period. It has the same pattern for the two sites, but University has greater positive values throughout the day (material absorbs and stores heat) and greater negative values at night (material releases stored heat). This demonstrates that the conventional rooftop more readily absorbs heat during the day and emits it at night compared to the polymer membrane on City Hall. For both sites, ΔQS is a cooling effect during sunlight hours (positive values) as heat is transported from the surface into deeper levels of the building materials. It remains positive, though minimal, at night at City Hall, but becomes a source of heat at University after 1700 LT, when values become negative and interior heat moves from the inner material toward the surface, raising surface temperature.
d. Sensible heat transport
Sensible heat transport is a function of the wind-dependent transport coefficient and the surface–air temperature differential, such that QH(t) = [S0 + S1υ(t)][TS(t) − TA(t)], with temperature in kelvins for the regression. Figure 11 illustrates the time-averaged diurnal variations in QH, wind speed, and [TS(t) − TA(t)]. City Hall has a positive QH value, indicating cooling, that is nearly a factor of 3 less than that of University, with values of 22.1 and 64.7 W m−2, respectively. The intercept S0 of the transport coefficient is greater at City Hall, suggesting that, all else (wind and surface–air temperature differential) being equal, sensible heat transport is greater at City Hall, perhaps because of effects such as heat island convection (Haeger-Eugensson and Holmer 1999; Morris et al. 2001), or the City Hall S0 excess may represent heat from one of the fluxes not explicitly derived (i.e., QF or ΔQA).
The difference in QH between the two sites is the net result of two opposing effects: the surface–air temperature differential and the transport coefficient. The diurnally averaged surface–air temperature differential affects QH more strongly than wind speed and is negligible at City Hall (0.8°C) as compared with University (3.6°C). Wind speed is lower at City Hall by 0.7 m s−1 as recorded in Table 4, which also contributes to lower QH. Decreased wind speed typically occurs with increased canopy size (Myrup et al. 1993) due to obstructions by rooftop structures and surrounding buildings. However, these limiting factors on QH at City Hall are partially compensated by a larger transport coefficient (S0), as shown in Table 2.
e. Air and surface temperatures
The 24-h diurnal patterns of TA and TS appear in Figs. 12a and 12b respectively, and Table 4 presents 24-h average values. Here, TA is measured 2 m above the underlying surface. Here, TS is calculated from εσTS(t)4 = L↑, where ε = 1. To determine site-specific effects on TS and TA, 24-h average values are first described and then diurnal patterns are examined to determine when and why one site becomes warmer or cooler than the other throughout the day.
Average air temperature is nearly identical for the two sites, being 26.8° and 26.9°C for University and City Hall, respectively. Average surface temperatures demonstrate a slightly greater difference, with values of 30.4° and 27.7°C for University and City Hall, respectively. This is an average TS differential of 2.7°C with University warmer than City Hall. This is primarily because of unobstructed K↓ on the conventional rooftop at University.
Figure 12a is a plot of TA throughout the 24-h period for both sites. Temperature differentials between the two sites occur during three distinct time periods. These periods are separated by intersections in TA when the sites have equal temperatures. Temperature intersections occur at 0830, 1500, and 2130 LT. Table 6 lists the average air temperatures for both sites for the time periods of 0830–1500, 1500–2130, and 2130–0830, and the maximum temperature differential between the sites and the time that it occurred. Between 0830 and 1500 TA is nearly identical for both sites. After urban sunset at City Hall, and before true sunset, between 1500 and 2130, University becomes an average of 0.6°C warmer and a maximum of 0.9°C warmer. After true sunset, City Hall has slower nocturnal cooling rates, so it becomes warmer than University by an average of 0.5°C, with a maximum ΔTA of 0.9°C just before sunrise.
Figure 12b is a plot of TS throughout the 24-h period for both sites. It illustrates two time periods when a temperature differential occurs between the sites. These time periods are separated by intersections in TS when the sites have equal temperatures. These intersections occur at 0730 and 2130. Table 7 lists the average surface temperatures for both sites for the time periods of 0730–2130 and 2130–0730, and the maximum temperature differential between the sites and the time that it occurred. University TS is 5.4°C warmer than that of City Hall between 0730 and 2130 because of direct solar irradiance on the University conventional rooftop. The maximum temperature differential is 12.4°C warmer at University, which occurs at 1510, just after urban sunset at City Hall. Slower nocturnal cooling rates at City Hall between 2130 and 0730 contribute to its surface temperature being an average of 1.1°C warmer than at University, and a maximum of 1.8°C warmer at 0520, just before sunrise.
f. Net radiation balance
This section presents a comparison of Q* and its components for the two sites (Table 3); first, average values are presented, and then diurnal patterns are presented. The Q* has a greater negative (incoming) value by ∼−40 W m−2 at the University than City Hall, with values of −93.3 and −49.2 W m−2, respectively. The K↓ primarily accounts for this, being ∼−100 W m−2 greater at University, with a value of −244.8 as compared with −146.9 W m−2 at City Hall. Heating by L↓ is ∼−30 W m−2 greater at City Hall than at University, with values of −419.5 and −389.8 W m−2, respectively. Obstructed K↓ and enhanced L↓ are both functions of the limited sky view at City Hall (Christen and Vogt 2004; Mills 1997b). Cooling values from K↑ are similar between University and City Hall, being 57.1 and 51.7 W m−2, respectively. Cooling from L↑ is greater at University by ∼20 W m−2, with a value of 484.2 as compared with 465.5 W m−2 at City Hall. Both ΔQS and QH compensate for the excess incoming Q* values, since QE is considered negligible. While ΔQS values at both sites are similar, being 28.6 W m−2 at University and 27.1 W m−2 at City Hall, cooling by QH is nearly a factor of 3 greater at University than City Hall, with values of 64.7 and 22.1 W m−2, respectively.
Individual flux components for the 24-h period are presented in Figs. 13 and 14 for University and City Hall, respectively. These figures illustrate the balance between the incoming (negative) and outgoing (positive) fluxes that act as warming and cooling effects, respectively, for the underlying surface. During periods of direct solar irradiance, (1 − α)K↓ and L↓ are dominant warming effects and are largely balanced by L↑, with QH providing the next largest cooling contribution, and ΔQS the remainder. After sunset, QH is a warming effect at both sites, while ΔQS is a warming effect at University but remains a cooling effect at City Hall.
5. Discussion
The originality of the model presented here lies in its ability to accurately calculate surface energy balance components for microscale (<103 m on a side) environments. The model does not require additional instrumentation such as a sonic anemometer or krypton hygrometer to calculate turbulent sensible or latent heat fluxes, respectively, which are sound methods (Grimmond and Oke 1999, 2002) but add to the cost and complexity of the application. Local scale (102–104 m) models also typically require tall towers on which to mount instruments in order to record measurements representative of a larger surface area.
The microclimate model presented here focuses on the individual structures or surfaces on which the instruments are installed. Unlike the local-scale urban meteorological parameterization scheme developed by Grimmond and Oke (2002), one of the most accurate and complete empirical models for local scale surface energy balance parameterizations, this model does not integrate individual surfaces and features into a characteristic neighborhood response; instead, it focuses on a surface area of approximately 1000 m2. The primary difference, therefore, is in its specificity.
Unlike modified vegetation or new urban canopy schemes, empirical models rely on statistical relations to derive values for energy balance components. Results are often very robust yet are limited to the range of conditions encountered in the original studies (Masson 2006). Empirical models are therefore very useful as validation tools for more complex schemes. In terms of direct applications, for example, a microclimate focus is beneficial for describing how a single structure or pedestrian walkway could be modified to mitigate surface heat; the reasons a cooling load may be too high for a particular building; or the benefits of a single green roof.
6. Conclusions
This paper presents a semiempirical energy balance model that does not require extensive instrumentation to produce robust values for microclimate energy fluxes. The model can be applied to a variety of environments and is not exclusively urban in its applicability. However, the microclimate length scales under consideration are similar to the dimensions of built structures and can therefore be especially useful in a variety of urban applications.
The intercomparison performed here demonstrates that the model accurately estimates surface energy fluxes that are influenced by site-specific features. The energy fluxes that show the greatest average difference between the two sites are K↓, L↓, and QH. These differences can be attributed to the structural characteristics of the microenvironments in which they occur. Surrounding high-rise structures and a rooftop wall render K↓ nearly 100 W m−2 less at City Hall than at University. Because of radiation from the walls of surrounding high-rise structures, which decreases cooling rates at night, L↓ is ∼30 W m−2 greater at City Hall. Cooling by QH is a factor of 3 less at City Hall because of diminished wind speed and a low surface-to-air-temperature differential. For the two sites, ΔQS is a similar cooling value when averaged over 24-h and has a similar pattern throughout the 24-h period, but University has higher maximum values in the daytime and lower minimum values at night. The QE is assumed negligible because of the lack of precipitation and surface liquid water, but under the particular conditions described here, its representation by the model cannot be evaluated.
Temperatures of built surfaces arise from a balance among the various energy fluxes that individually act to heat or cool an area. The demand for energy to heat or cool interior building space depends on both outdoor air temperatures and the temperatures of built surfaces, specifically the roofs and walls of buildings exposed to the atmosphere. The temperature of a solid surface in general differs from the air temperature and depends on properties of the structure such as its albedo, thermal conductivity, heat capacity, and the availability of liquid water for evaporation. In addition, the external surroundings of a structure can influence its surface temperature. All of these factors produce microclimates, exemplified by the City Hall and University sites studied here. Although these locations are in close geographic proximity, they experience somewhat different diurnal cycles in surface temperature and would, a priori, have different requirements for interior temperature control.
Acknowledgments
Although the research described in this article has been funded wholly or in part by the U.S. Environmental Protection Agency through Grant Number CR-83089001-0 to the University of Chicago, it has not been subjected to the agency’s required peer and policy review and therefore does not necessarily reflect the views of the agency and no official endorsement should be inferred. The authors thank Kimberly Worthington, Chicago Department of Environment; John Albrecht, Chicago Department of General Services; and Roger McGinty and Wally Gorman, Chicago City Hall Building Engineers. Three anonymous reviewers provided valuable comments that contributed to this paper. Aerial images were produced with Google Earth.
REFERENCES
Akbari, H., S. Konopacki, and M. Pomerantz, 1999: Cooling energy savings potential of reflective roofs for residential and commercial buildings in the United States. Energy, 24 , 391–407.
Akbari, H., M. Pomerantz, and H. Taha, 2001: Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas. Sol. Energy, 70 , 295–310.
Arnfield, A. J., and C. S. B. Grimmond, 1998: An urban canyon energy budget model and its application to urban storage heat flux modeling. Energy Build., 27 , 61–68.
Barzyk, T. M., 2006: Climatic effects of urbanization: Urban, suburban, and rural microclimates in the Chicago region. Ph.D. dissertation, University of Chicago, 173 pp. [Available from ProQuest Dissertations and Theses, 300 N. Zeeb Rd., Ann Arbor, MI 48106-1346.].
Bretz, S., H. Akbari, and A. Rosenfeld, 1998: Practical issues for using solar-reflective materials to mitigate urban heat islands. Atmos. Environ., 32 , 95–101.
Christen, A., and R. Vogt, 2004: Energy and radiation balance of a central European city. Int. J. Climatol., 24 , 1395–1421.
De Schiller, S., and J. M. Evans, 1996: Training architects and planners to design with urban microclimates. Atmos. Environ., 30 , 449–454.
Evans, J. M., and S. De Schiller, 1996: Application of microclimate studies in town planning: A new capital city, an existing urban district and urban river front development. Atmos. Environ., 30 , 361–364.
Golany, G. S., 1996: Urban design morphology and thermal performance. Atmos. Environ., 30 , 455–465.
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. Meteor., 38 , 922–940.
Grimmond, C. S. B., and T. R. Oke, 2002: Turbulent heat fluxes in urban areas: Observations and a local-scale urban meteorological parameterization scheme (LUMPS). J. Appl. Meteor., 41 , 792–810.
Grimmond, C. S. B., S. K. Potter, H. N. Zutter, and C. Souch, 2001: Rapid methods to estimate sky-view factors applied to urban areas. Int. J. Climatol., 21 , 903–913.
Grimmond, C. S. B., J. A. Salmond, T. R. Oke, B. Offerle, and A. Lemonsu, 2004: Flux and turbulence measurements at a densely built-up site in Marseille: Heat, mass (water and carbon dioxide), and momentum. J. Geophys. Res., 109 .D24101, doi:10.1029/2004JD004936.
Haeger-Eugensson, M., and B. Holmer, 1999: Advection caused by the urban heat island circulation as a regulating factor on the nocturnal urban heat island. Int. J. Climatol., 19 , 975–988.
Kanda, M., T. Kawai, and K. Nakagawa, 2005: A simple theoretical radiation scheme for regular building arrays. Bound.-Layer Meteor., 114 , 71–90.
Lowry, W. P., 1977: Empirical estimation of urban effects on climate: A problem analysis. J. Appl. Meteor., 16 , 129–135.
Masson, V., 2006: Urban surface modeling and the meso-scale impact of cities. Theor. Appl. Climatol., 84 , 35–45.
Masson, V., C. S. B. Grimmond, and T. R. Oke, 2002: Evaluation of the Town Energy Balance (TEB) scheme with direct measurements from dry districts in two cities. J. Appl. Meteor., 41 , 1011–1026.
Mills, G., 1997a: An urban canopy-layer climate model. Theor. Appl. Climatol., 57 , 229–244.
Mills, G., 1997b: The radiative effects of building groups on single structures. Energy Build., 25 , 51–61.
Morris, C. J. G., I. Simmonds, and N. Plummer, 2001: Quantification of the influences of wind and cloud on the nocturnal urban heat island of a large city. J. Appl. Meteor., 40 , 169–182.
Myrup, L. O., C. E. McGinn, and R. G. Flocchini, 1993: An analysis of microclimatic variation in a suburban environment. Atmos. Environ., 27B , 129–156.
Offerle, B., C. S. B. Grimmond, and T. R. Oke, 2003: Parameterization of net all-wave radiation for urban areas. J. Appl. Meteor., 42 , 1157–1173.
Oke, T. R., 1988: The urban energy balance. Prog. Phys. Geogr., 12 , 471–508.
Oke, T. R., 2004a: Siting and exposure of meteorological instruments at urban sites. Preprints, 27th NATO/CCMS Int. Tech. Meeting on Air Pollution Modelling and Its Application, Banff, AB, Canada, NATO/CCMS, 1–14.
Oke, T. R., 2004b: Initial guidance to obtain representative meteorological observations at urban sites. Instruments and Observing Methods Rep. 81, WMO-TD 1250, 51 pp.
Pearlmutter, D., and P. Berliner, 2005: Evaluation of urban surface energy fluxes using an open-air scale model. J. Appl. Meteor., 44 , 532–545.
Pearlmutter, D., A. Bitan, and P. Berliner, 1999: Microclimatic analysis of “compact” urban canyons in an arid zone. Atmos. Environ., 33 , 4143–4150.
Santamouris, M., N. Papanikolaou, I. Livada, I. Koronakis, C. Georgakis, A. Argiriou, and D. N. Assimakopoulos, 2001: On the impact of urban climate on the energy consumption of buildings. Sol Energy, 70 , 201–216.
Stewart, J. B., W. P. Kustas, K. S. Humes, W. D. Nichols, M. S. Moran, and H. A. R. De Bruin, 1994: Sensible heat flux-radiometric surface temperature relationship for eight semiarid areas. J. Appl. Meteor., 33 , 1110–1117.
Sun, J., and L. Mahrt, 1995: Relationship of surface heat flux to microscale temperature variations: Application to BOREAS. Bound.-Layer Meteor., 76 , 291–301.
Aerial view of the University and City Hall sampling locations, 10.6 km apart. The picture is oriented with north at top.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Aerial view of the City Hall sampling location with 3D rendering of buildings. The picture is oriented with north on left.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Close-up view of the City Hall sampling location. The adjoining black tar rooftop can be seen on the right; at the top is edge down to courtyard; an ∼5-m-tall brick wall is on the left. North is oriented toward the top of the picture. In this picture, the radiometer is oriented sideways; for the results presented here, it was oriented perpendicular to the roof surface.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Aerial view of the University sampling location. The picture is oriented with north on left.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Close-up view of the University sampling location. North is oriented toward the bottom right-hand corner.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Instrument intercomparison graphs based on more than 2000 points. (a) Air temperature (TA); (b) wind speed (υ); (c) net solar (KNET); (d) net thermal (LNET). City Hall instruments are plotted on the x axis, and University on the y axis, collocated on University roof from 21 Sep to 5 Oct 2005 at 10-min intervals. Regression equations are listed in Table 1. No data corrections are required to account for instrument nonequivalence. Pattern in (b) is because the recording of wind speed is as an integer value, such that the middle points are 1:1, and directly above and below are y = x + 0.45 m s−1 and y = x − 0.45 m s−1 (equivalent to 1-mph discrepancy), respectively.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Diagram of the energy balance model geometry. Surface with depth “D” is assumed to respond instantaneously to atmospheric interface energy fluxes. Surface at x = −L could be interior of building or some depth underground. Positive energy fluxes flow from surface (x = 0) to atmosphere (x > 0); negative ones flow from depth D to x = −L.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
(a) Incoming solar shortwave radiation (K↓) and (b) reflected shortwave (αK↓) for the City Hall and University sampling locations averaged into a 24-h period beginning at 0400 LT (∼2 h before dawn) from 7 days of results between 29 Jul and 5 Aug 2005.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Incoming thermal longwave radiation (L↓) for the City Hall and University sampling locations averaged into a 24-h period beginning at 0400 (∼2 h before dawn) from 7 days of results between 29 Jul and 5 Aug 2005.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Net effect of evaporation and conduction (QE −ΔQS) for the City Hall and University sampling locations averaged into a 24-h period beginning at 0400 (∼2 h before dawn) from 7 days of results between 29 Jul and 5 Aug 2005.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
(a) Sensible heat transport (QH), (b) wind speed (υ), and (c) surface–air temperature differential (TS − TA) for the City Hall and University sampling locations averaged into a 24-h period beginning at 0400 (∼2 h before dawn) from 7 days of results between 29 Jul and 5 Aug 2005.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
(a) Air (TA) and (b) surface (TS) temperatures for the City Hall and University sampling locations averaged into a 24-h period beginning at 0400 (∼2 h before dawn) from 7 days of results between 29 Jul and 5 Aug 2005.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Energy balance components for University. Negative values indicate a warming effect; positive values are for a cooling effect. Shown is the average of 24 h beginning at 0400 (∼2 h before dawn) from 7 days of results between 29 Jul and 5 Aug 2005. The dashed line at 0 axis is for reference.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
As in Fig. 13, but for energy balance components for City Hall.
Citation: Journal of Applied Meteorology and Climatology 47, 3; 10.1175/2007JAMC1431.1
Regression equations and associated R2 values that result from plotting collocated measurements from the radiometers and weather stations used at University (y variable) and City Hall (x variable) from the period from 21 Sep to 5 Oct 2005 (>2000 points) on University rooftop.
Energy fluxes averaged (denoted by angle brackets) over 7 consecutive 24-h periods: incoming and reflected shortwave solar (〈−K〉, 〈K〉), incoming and upwelling thermal longwave [〈−L〉, 〈εσTS(t)〉], net radiation (〈Q*〉), the net effect of evaporation and conduction (〈QE − ΔQS〉), and sensible heat transport (〈QH〉). A negative number indicates a net heating of the solid surface. A positive number corresponds to a net cooling of the surface.
Atmospheric variables averaged (denoted by brackets) over the 7-day sampling period: surface and air temperatures (〈TS〉, 〈TA〉) and wind speed (〈υ〉).
Mean daily values of air temperature (〈TA〉), incoming solar radiation (〈K↓〉), and wind speed (〈υ〉) for each of the 7 days, 29 Jul to 5 Aug 2005. Days 1–6 had mostly clear to partly cloudy conditions; day 7 was mostly cloudy. No precipitation occurred during the week.
Average air temperatures (TA) for the three time periods that occur between intersections of the temperature curves in Fig. 12a.
Average surface temperatures (TS) for two time periods that occur between intersections of the temperature curves in Fig. 12b.
