- Email: [email protected]
Assessing the effects of landscape design parameters on intra-urban air temperature variability: The case of Beijing, China
Building and Environment 76 (2014) 44e53
Contents lists available at ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
Assessing the effects of landscape design parameters on intra-urban air temperature variability: The case of Beijing, China Hai Yan, Shuxin Fan, Chenxiao Guo, Fan Wu, Nan Zhang, Li Dong* College of Landscape Architecture, Beijing Forestry University, Beijing 100083, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 19 January 2014 Received in revised form 5 March 2014 Accepted 7 March 2014
Understanding the causes of the intra-urban air temperature variability is a first step in improving urban landscape design to ameliorate urban thermal environment. Here we investigated the spatial and temporal variations of air temperature at a local scale in Beijing, and their relationships with three categories of landscape design parameters, including the land cover features, site geometry, and spatial location. Air temperature measurements were conducted during the winter of 2012 and the summer of 2013 by mobile traverses. The results showed that spatial temperature difference between the maximum and minimum observed temperature in the study area ranged from 1.2 to 7.0 C, and varied depending on season and time of the day. The magnitude and spatial characteristic of the air temperature variations depend strongly on the landscape parameters characterizing the immediate environment of the measurement sites. Increasing the percentage vegetation cover could significantly decrease air temperature, while the increase of building area would significantly increase it. In addition, the observed air temperature increased as the sky view factor (SVF) increased during daytime, while a contrary tendency was observed during nighttime. However, the impacts of SVF on air temperature were context-dependent. Furthermore, the air temperature increased with increasing distance from the park and water body boundary. Our findings also indicated that the relative importance of these landscape parameters in explaining air temperature differences varied among different times and seasons. Therefore, if appropriately combined, all investigated landscape parameters can effectively improve urban thermal environment on a yearly basis. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Urban heat island Air temperature Landscape design Land cover Vegetation Sky view factor
1. Introduction Growing urban populations and urban expansion, with increasing built-up areas and human activities, results in significant modifications in the underlying surface properties and energy balance, and thus alters the urban climate [1]. A distinct feature of urban climate is the urban heat island (UHI) effect, which is directly related to the conversion of land cover from rural to urban covers [2e4]. Also, intra-urban surface cover variations result in the preferred spatial clustering of urban canopy layer (UCL) scale surfaces and surface units. This results in distinctive surface energy balance and microclimatic characteristics at a local scale [5]. Given the vast array of surface that forms the urban mosaic-like structure, it is not surprising that the microclimate within the urban environment varies significantly within the micro to local
* Corresponding author. Tel./fax: þ86 10 62336605. E-mail address: [email protected] (L. Dong). http://dx.doi.org/10.1016/j.buildenv.2014.03.007 0360-1323/Ó 2014 Elsevier Ltd. All rights reserved.
scale [6,7]. Urban microclimate both influences and is influenced by human behavior and decision-making, due to the complex interactions between air temperature and microscale landscape parameters [8]. Therefore, better understanding the causes of this intra-urban air temperature variability is critical in improving landscape design strategies to ameliorate urban thermal environment [9e11]. The first category of landscape parameters that influence urban microclimate is the land cover features. As urban areas develop and density increases, impervious urban surface covers replace permeable natural vegetation covers. Finally, impervious surface change the reflectivity and energy balance of surfaces resulting in locally higher air temperature. For this reason, vegetated areas play a critical role in moderating and dampening the warming effects of impervious land covers in urban environments [12e14]. Many field-based measurements have found that urban green areas are generally cooler than their surrounding built-up areas, and can produce air temperature differences up to 1e7 C, a phenomenon referred to as a “park cool island” [15e21]. Furthermore, a few
H. Yan et al. / Building and Environment 76 (2014) 44e53
qualitative associations between air temperature and its surrounding land cover features have been reported [22e24]. Sun [23], for example, showed that air temperature significantly correlated with green ratio and building ratio at night in Taichung City, Taiwan. Yokobori and Ohta [24] investigated the effect of land cover on ambient air temperatures and found that air temperatures varied significantly according to ambient land cover types, and air temperatures decreased as the amount of vegetated area around the sites increased. Site geometry is another crucial landscape parameter because of its importance in determining the receipt and loss of radiation. An important measure of surface geometry of a given site is the sky view factor (SVF), i.e. the fraction of the overlying hemisphere occupied by the sky [25]. During the day, the lower SVF at the canyon will result in less incoming solar radiation, and thus has an important effect on the ground surface temperature as well as the air temperature directly above it. The “urban cool island” phenomenon documented in many previous studies is mainly related to the low SVF [26,27]. During nighttime, the decreased SVF below roof level reduces the loss of long-wave radiation to the sky and also reduced turbulent heat transfer in the often calm canyon air. Therefore, theoretically it is considered to be a major factor of the nocturnal UHI phenomenon [28e31]. However, some studies have also found that SVF does not have a large importance on city temperature, especially not air temperature, and that it should not be assigned to much importance [32e34]. From past research it remains somewhat unclear the role that site geometry plays in its effects on urban air temperature. In addition, there can be no doubt about the important role of heat advection between contrasting surfaces at the microscale in the UCL [35]. Hence, the air temperature at a site is also affected by dispersion through turbulence and advection from its surroundings. In particular, vegetated areas or water bodies within the urban environment are moist and cool compared to their surrounding built areas [36], and thus the cooling effects of vegetation or water could extend into its surrounding area. For example, Ca et al. [37] investigated the influence of a park on the urban summer climate in the nearby areas, and found at noon, the park can reduce by up to 1.5 C the air temperature in a busy commercial area 1 km downwind. Observations made by other researches also show an extension of the cooling effect [16,34,38e40]. As this brief overview shows, the landscape design parameters of land cover, urban geometry, and the spatial location (proximity to the park or water) have been found to influence the local air temperature. The problem, however, is that different researchers look at the problem form a different angle using different landscape parameters, and it is very difficult to conclude which particular landscape factors would be more important in effecting the air temperature within the urban context. In addition, urban climate is also affected by external influences such as the topographic features, season and prevailing weather. Therefore, it is important to control geographical, seasonal and meteorological (i.e. wind speed and cloud cover) variables as much as possible to understand the location specific impacts on changes in urban air temperature. In this study, we investigate the spatial and temporal differences of air temperature at a local scale in Beijing, and to gain insights on its linkage to the landscape parameters including percentage of building area, percentage vegetation cover, SVF, distance to park, and distance to water body. Specifically, this study tried to answer the following two questions: (1) how do the different landscape parameters differentially contribute to explaining the variance of local air temperatures, and (2) if the relative importance of these landscape parameters in explaining local air temperature varies by season and time of day?
45
2. Methodology 2.1. Study area and measurement sites Beijing (39 560 N, 116 200 E), the capital of China, is located in the northern part of the North China Plain. It is the second largest city in China with a total population of 19.6 million by the end of 2010. It has a monsoon influenced humid continental climate characterized by hot and humid summers and generally cold, windy and dry winters. According to the climatological normals (1971e2000), January is the coldest month with an average temperature of 3.7 C, while July is the hottest month with an average temperature of 26.2 C. The highest temperature in 1971e2000 was 41.9 C, and the coldest temperature was 18.3 C. The main wind direction is from southeast to northeast in summer and in reverse during winter. Since 1978, Beijing’s urban population and yearly construction area have been gradually increasing. This urbanization process, with its increasing built-up areas and anthropogenic activities, results in significant modifications in the underlying surface properties and the quick increase in the intensification of the UHI effect [41]. The measurements took place at the Beijing Olympic Park and its surrounding built-up areas in the Beijing city (Fig. 1). The study area is rectangular, 3.2 2.1 km in size and situated at the northern part of the city. It can be regarded as a typical area in the process of urbanization in Beijing. This area is very flat, and thus temperature difference due to topographic influence can be ignored. In addition, this district has a perfectly orthogonal urban geometry and the streets are oriented EeW and NeS, with an average width of 25 m. Therefore, it is a good study area to explore the relationship between air temperature and landscape design factors. In this study area, 26 measurement sites were chosen with the intention to investigate the quantitative relationship between air temperature and landscape variables (Fig. 1). A description of the measurement sites is given in Table 1. These measurement sites located sufficiently close to one another to be affected by uniform meso-scale climate conditions, and yet, also affected by distinctly different micro-scale landscape characteristics. The nature of the adjacent underlying surface greatly affects air temperature. Thus, to standardize temperature, all air temperature measurements were made over a common surface: paved road, as shown in Fig. 2(a). 2.2. Air temperature measurements Mobile traverses were used to conduct air temperature measurements on 8 days during the winter of 2012 (December 2012 to February 2013) and the summer of 2013 (June to August). Survey times were early afternoon (14:00e15:00) and midnight (23:00e 00:00). All observations were carried out during clear and windless days, i.e. when the UHI effects could be expected to be most pronounced. Avoidance of windy and cloudy conditions also minimizes the influence of meteorological variables. Measurements of air temperature were taken using a mobile station with a thermistor temperature sensor connected to a data logger with a precision of 0.1 C along the measurement route. The temperature sensor was fitted within a radiation shield, which mounted onto the front of a bicycle at 1.5 m above ground level. A GPS travel recorder was used to note geographical coordinates of latitude, longitude, and altitude synchronized to observations. Three fixed weather stations were installed in different regions within the study area for recording of the daily trend of air temperature. These data were used to correct the values taken with the mobile station because the measurements at different sites were not instantaneous. Each traverse took about 1 h to complete, and all data were adjusted to the beginning time of each transect by
46
H. Yan et al. / Building and Environment 76 (2014) 44e53
Fig. 1. Location of the study area and measurement sites. The numbers represent the mobile measurement sites, while the lowercase letters represent the three fixed air temperature measurement sites.
modifying values using temperature change rates over the study area based on fixed weather station measurements of air temperature. Wind speed and direction were obtained from the weather observatory station near the measurement sites. In most cases, the wind does not change significantly during the individual one-hour measurement session. 2.3. Measurement and computations of landscape parameters Numerous parameters have been developed to assess and quantify the effects of urban environment characteristics on air temperature. For this study, three categories of landscape parameters including land cover feature, site geometry and spatial location, were selected to measure site environmental characteristics. These parameters were selected based on the following principles: (1) the potential effects on air temperature, (2) easily calculated, (3) readily understood, easily controlled by design professionals, and (4) minimal redundancy. Land cover features included the percentage of building area and percentage vegetation cover. Site geometry was measured using SVF and spatial location was measured using distance to park and distance to water body. Aerial photographs (Google Maps, September 15, 2012) were used to acquire the land cover features in the study area. Since the
variation in air temperature with regard to the land cover composition of each measurement sites need to be quantified, a buffer zone was used to accomplish this. According to the results of some previous studies, air temperature in an urban location is mainly affected by its surrounding surface cover with a few hundred meters in radius, and the far field tends to have negligible impacts due to the intricate urban fabrics [22,24]. Thus, a buffer zone with 150 m in radius was used in our study. Using a similar method to Krüger and Givoni [22], the percentage of building area and vegetation cover for each measurement site were calculated with AutoCAD, by drawing corresponding areas on the aerial photograph (Fig. 2(b)). For the determination of the SVF, fisheye images at each measurement site were taken by a fisheye lens (Sigma 8 mm circular fisheye lens) coupled to a Cannon EOS 5D Mark II digital camera. The fisheye images were processed into black and white image, in which the sky is white and the building and trees are black. Then, the images were put into RayMan software for the SVF calculation. The SVF were calculated separately for summer and winter because of the seasonal variation in vegetation, as shown in Fig. 2(c and d). The effect of park and water body has been estimated based on the straight distance between each measurement site and the edge of the park and the water body. In the present study, the park refers mainly to the Beijing Olympic Forestry Park, and the water body refers mainly to the Dragon-shaped Olympic Water System in
H. Yan et al. / Building and Environment 76 (2014) 44e53
47
Table 1 Landscape characteristics and description of measurement sites. Site number
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Landscape parameters
Description of measurement sites
BA (%)
VC (%)
SVF (summer)
SVF (winter)
DP (m)
DW (m)
23.4 28.6 29.7 32.1 31.9 7.3 7.1 0.0 4.2 9.5 13.2 0.6 11.2 17.8 28.9 28.7 31.4 24.7 12.3 21.6 10.3 32.1 32.8 12.4 1.3 10.3
33.1 31.3 27.8 27.4 25.4 53.0 57.6 82.1 62.2 54.7 20.1 44.8 65.9 20.1 27.5 15.7 39.7 34.6 36.7 16.6 27.9 19.8 30.2 42.7 71.8 61.0
0.63 0.62 0.59 0.66 0.49 0.70 0.49 0.66 0.72 0.76 0.39 0.76 0.58 0.61 0.59 0.32 0.31 0.11 0.59 0.46 0.76 0.51 0.38 0.38 0.46 0.46
0.67 0.71 0.60 0.67 0.60 0.80 0.75 0.82 0.79 0.79 0.64 0.83 0.77 0.60 0.60 0.47 0.51 0.55 0.75 0.46 0.76 0.50 0.45 0.49 0.75 0.60
180 660 1259 1375 1084 612 128 5 180 626 1019 1259 1004 613 505 625 1006 1242 1340 918 907 828 204 44 5 45
702 1097 1202 949 606 619 458 321 292 128 18 212 396 410 695 1004 1010 639 282 919 90 722 870 452 35 437
Broad street, multi-storey buildings Broad street, different-sized buildings High-rise and multi-storey buildings Broad street, multi-storey buildings Broad street, high-rise buildings Park’s perimeter road, tree covered Park’s perimeter road, tree covered Inside park, open grass area Inside park, open paved area Inside park, open paved area Waterside, tree covered Open area, parking lot Broad street, street trees Broad street, high-rise buildings High-rise buildings, tree covered High-rise buildings, tree covered High-rise buildings, tree covered Multi-storey buildings, tree covered Broad street, high-rise buildings High-rise building, heavy traffic Open area High-rise buildings, heavy traffic High-rise buildings, tree covered High-rise buildings, park edge Inside park, open vegetated area Multi-story building, park edge
Note: BA, VC, SVF, DP, and DW refer to the percentage of building area, percentage vegetation cover, sky view factor, distance to park, and distance to water body, respectively.
Beijing City. The detailed statistics of these landscape parameters are listed in Table 1. 2.4. Statistical analyses We first performed simple linear regression to analyze the bivariate relationship between air temperatures and percentage of building area, percentage vegetation cover, SVF, distance to park and distance to water body. Further, a multivariate regression analysis was used to quantify the relative contribution of five landscape variables to the air temperature differences. The predictive model was assumed as:
Y ¼ a þ b1 BA þ b2 VA þ b3 SVF þ b4 DP þ b5 DW
(1)
Where Y represents the value of air temperature ( C); a is a constant; b1, b2.b5 are the coefficients for each independent variable; BA (%) is the percentage of building area; VA (%) is the percentage vegetation area; SVF is the sky view factor; DP (m) is the distance to park; and DW (m) is the distance to water body. All of the statistical analyses were performed using SPSS 17.0 software. 3. Results 3.1. Temporal and spatial variation of air temperatures The spatial patterns of the average temperature of four mobile traverses among different times and seasons within the study area are presented in Fig. 3. The air temperature differences among different locations are very significant. During daytime in summer (Fig. 3(a)), the hottest place was at site 4 with mean air temperature reaching 35.5 C. This site is located in a shallow street canyon, which is fully exposed to solar radiation during daytime. The continuous heating of the pavement surface by solar radiation and the release of anthropogenic heat contributes to the high air
temperature here. It is worth noting that the air temperatures measured at sites 16 and 17, which are located in the center of the commercial area, were not much higher even with the expected high anthropogenic heat flux. This is because the high-rise buildings located on both sides of the street and mature street trees can provide very good shading. At this time, the coolest sites were observed at sites 8 and 25, both of which were located inside the park with dense vegetation cover. Even though these sites were exposed to solar radiation during the daytime, evapotranspiration from the grass surface and trees caused low air temperature. During nighttime in summer (Fig. 3(b)), the pattern of temperature distribution in the study area is very clear. The park area now became a cool island surrounded by hot areas. The mean air temperature difference between the lowest inside the park and the highest in the building area was 3.2 C. During daytime in winter, the air temperatures distribution profile is rather complicated, as shown in Fig. 3(c). At this time, sites 1, 14 and 19 were found to display a higher air temperature than other sites. The probable reasons are that they are exposed to the open space, a reduction of building shade at their immediate environment and with increased exposure to direct sunshine. The air temperature at site 24 was the lowest as it is located at the south-side of the EeW oriented street canyon, well shaded by highrise apartment buildings. During night in winter (Fig. 3(d)), the hottest site was site 22, which is located in a heavy traffic street characterized by low ventilation, high business activities, multistoried buildings and heavy road traffic. The coldest site was observed at the north-central part of the park, where mean air temperature reached 8.5 C. This open space stores less heat during daytime and releases the heat during nighttime at a faster rate compared to densely built environment. To analyze the temporal variation of temperature, DTmax-min, which is the difference between the maximum and minimum observed temperature, was calculated for each measurement. Fig. 4 clearly shows that the differences of the DTmax-min among different
48
H. Yan et al. / Building and Environment 76 (2014) 44e53
Fig. 2. Landscape characteristics of one example of the measurement sites (site 1). (a) measurement environment, (b) land cover features, (c) SVF in summer, and (d) SVF in winter.
times and seasons are very significant. In summer, the daytime DTmax-min ranged from 1.7 to 1.9 C, and it ranged from 2.1 to 4.5 C during nighttime. In winter, the daytime DTmax-min ranged from 1.2 to 2.1 C, and the nighttime DTmax-min ranged from 4.3 to 7.0 C. On average, the DTmax-min was largest in winter nighttime (5.9 C), followed by summer nighttime (3.4 C), summer daytime (1.8 C), and winter daytime (1.7 C). To explore whether the driving mechanisms of air temperature differences varied with times and seasons, we further analyze the relationship of air temperature between daytime and nighttime, and summer and winter. From the aspect of diurnal pattern of spatial difference in air temperature, nighttime air temperature increased significantly with daytime air temperature in summer
(Fig. 5(a); R2 ¼ 0.45, P < 0.001). However, there was no relationship between daytime and nighttime air temperature in winter (Fig. 5(b); R2 ¼ 0.00, P ¼ 0.990). For seasonal pattern of spatial difference in air temperature, there was no relationship between summer and winter daytime air temperature (Fig. 5(c); R2 ¼ 0.08, P ¼ 0.154), while there was a significant positive relationship between summer and winter nighttime air temperature (Fig. 5(d); R2 ¼ 0.75, P < 0.001). 3.2. Influence of the land cover features on air temperature The effects of the composition of land cover features on air temperatures vary clearly among different times and seasons, as
Fig. 3. Spatial variation of average air temperatures (average of four measurement days) by observation site at (a) summer daytime, (b) summer nighttime, (c) winter daytime, and (d) winter nighttime.
H. Yan et al. / Building and Environment 76 (2014) 44e53
49
building area (Fig. 6(c); R2 ¼ 0.664, P < 0.001) and a strong negative relationship between air temperature and the percentage vegetation cover (Fig. 6(d); R2 ¼ 0.620, P < 0.001). 3.3. Influence of the site geometry on air temperature
Fig. 4. Box plots showing differences between the warmest and coldest observed temperature among different times and seasons. The bulging dots in the middle indicate the median, the boxes indicate the upper and lower quantiles, and the whiskers indicate the range.
shown in Fig. 6. During daytime in summer, there was a weak positive relationship between air temperature and the percentage of building area (Fig. 6(a); R2 ¼ 0.256, P ¼ 0.008) and a weak relationship between air temperature and the percentage vegetation cover (Fig. 6(b); R2 ¼ 0.202, P ¼ 0.021). During nighttime in summer, there was a strong positive relationship between air temperature and the percentage of building area (Fig. 6(a); R2 ¼ 0.674, P < 0.001) and a strong negative relationship between air temperature and the percentage vegetation cover (Fig. 6(b); R2 ¼ 0.702, P < 0.001). During daytime in winter, there was no significant relationship between air temperature and both of percentage of building area and percentage vegetation cover (Fig. 6(c and d)). During nighttime in winter, there was a strong positive relationship between air temperature and the percentage of
During daytime in both summer and winter, air temperatures increased with increasing SVF (Fig. 7), suggesting that during daytime, when there was solar radiation, the lower the sky openness, the lower the air temperature was. However, the simple linear regression analysis did not show a statistical significant relationship between air temperature and SVF in both summer daytime (Fig. 7(a); R2 ¼ 0.100, P ¼ 0.115) and winter daytime (Fig. 7(b); R2 ¼ 0.002, P ¼ 0.811). During nighttime, it is clearly shown that the correlation has changed into negative, which means that during this time, the net outgoing long-wave decreases at location that has a low SVF values, resulting in higher temperatures. According to the obtained statistical measurements, about 45.2% (P < 0.001) of the intra-urban variations in air temperature can be explained by the variation in SVF during winter nighttime (Fig. 7(b)). SVF, however, does not explain summer nighttime air temperature variations (Fig. 7(a); R2 ¼ 0.048, P ¼ 0.281). 3.4. Influence of the distance to park and water on air temperature Fig. 8 shows the relationships between air temperature and distance to park and distance to water body. During daytime and nighttime in summer and winter, air temperature increased with increasing distance to park, but more rapidly at nighttime than at daytime and more rapidly at winter nighttime than at summer nighttime. The R2 were 0.248, 0.327, and 0.486 during summer daytime, summer nighttime, and winter nighttime, respectively, indicating that the variable (distance to park) account for 24.8%, 32.7%, and 48.6% of variance in the air temperature distribution. However, there was no significant relationship between air temperature and distance to park during daytime in winter. The air temperatures determined by the variable of distance to water was
Fig. 5. Relationships between (a) summer daytime temperature and summer nighttime temperature, (b) winter daytime temperature and winter nighttime temperature, (c) summer daytime temperature and winter daytime temperature, and (d) summer nighttime temperature and winter nighttime temperature.
50
H. Yan et al. / Building and Environment 76 (2014) 44e53
Fig. 6. Relationships between air temperature and land cover features for the percentage of building area and percentage vegetation cover in summer (a, b) and in winter (c, d).
very similar to that determined by the distance to park. The R2 were 0.149, 0.471, and 0.462 during summer daytime, summer nighttime, and winter nighttime, respectively, and there was also no significant relationship between air temperature and distance to water at winter daytime. 3.5. Multiple regression analysis Multiple regression analysis was carried out in order to find out how well the observed air temperature differences could be explained by the combination of the five landscape variables (Table 2). The regression results offer insights into the influence of different variables on air temperature among different times and seasons. The coefficient of determination (R2) represented the proportion of the variation in air temperature could be explained by the regression models; and the standardized coefficients (Beta coefficients) of predictive models represented the relative contributions of different landscape variables to the air temperature difference. During daytime in summer, approximately 64.5% of the variation in air temperature was explained jointly by the five landscape variables. The only significant factor in this model was SVF. The comparison of beta coefficients indicates that the SVF was also the most important variable. According to this model, for every 10% increase in SVF will increase air temperature by 0.14 C. During nighttime in summer, about 86.7% of the variation in air temperature was explained jointly by the five landscape variables. The only significant factor in this model was percentage vegetation cover. A 10% increase in vegetation cover decreased air temperature by 0.23 C. During daytime in winter, approximately 30.3% of the variation in air temperature was explained jointly by the five landscape variables. The only significant factor in this model was SVF. The comparison of beta coefficients indicates that the SVF was also the most important variable. According to this model, for every 10% increase in SVF will increase air temperature by 0.24 C. During nighttime in summer, about 86.7% of the variation in air temperature was explained jointly by the five landscape variables. The significant independent variables were the percentage
vegetation cover, SVF, distance to park and distance to water. The negative coefficients of percentage vegetation cover and SVF suggest that an increase in the percentage vegetation cover and SVF will decrease air temperature. In contrast, the positive coefficients of distance to park and water indicate that temperature will increase with the increase of distance from park and water body boundary. 4. Discussion Results from the surveys demonstrate large spatial differences in temperatures in the study area. The magnitude and spatial characteristic of the differences varied depending on season and time of the day. At nighttime, the pattern of temperature distribution in the study area was very clear. The park area became a cool island surrounded by hot built areas. However, the spatial pattern by day tended to be less well defined. It is also worth noting that at this time, air temperature at high-rise building areas might be cooler than the sites inside the park, resulting in an urban cool island. This phenomenon has also been observed by Chow and Roth [42], mainly because of the slower warming of urban surfaces due to the solar heat storage of building materials and the shading effect of the building and trees. The maximum air temperature differences between the warmest and coldest site were largest in winter nighttime, followed by summer nighttime, summer daytime, and winter daytime. These basic patterns are generally similar to many previous urban heat island and park cool island studies conducted in cities of various sizes. Air temperature differences were greater at night, especially in winter during clear, calm weather conditions. This is largely attributable to the difference in radiative cooling rates between natural vegetation and built area. The vegetated area was more exposed to the sky than the building area, which allows for a higher cooling rate of this area. At the same time, the decrease in temperature in the building area was lower because this area is surrounded by buildings, which influences the loss of long-wave radiative to the sky. Since the difference in radiative cooling rates between the vegetated area and the building area was particularly pronounced in winter, when the canopies are leafless and at the
H. Yan et al. / Building and Environment 76 (2014) 44e53
51
Fig. 7. Relationships between air temperature and SVF in summer (a) and in winter (b).
Fig. 8. Relationships between air temperature and distance to park and distance to water body in summer (a, b) and in winter (c, d).
same time as lower soil moisture leads to low thermal heat capacity in the vegetated area, a maximum temperature difference of 7.0 C occurred during nighttime in winter. We further analyzed the correlation of mean air temperature pattern between different times and seasons. There was a significant relationship between nighttime air temperature and daytime air temperature in summer, which suggests that there is inertia of
daytime processed into the following night. However, we found no relationship between daytime and nighttime air temperature in winter, which means that the factors driving air temperature variations during the day are different than those at night in winter. We also found that there was no relationship between summer and winter daytime air temperature, while there was a significant positive relationship between summer and winter nighttime air
Table 2 Regression results of air temperature and five landscape variables. The bold figures are significant variable with p < 0.05. Variables
Summer noon
Summer night
Coefficient
BA VC SVF DP DW Constant R2 Adjusted R2
Sig.
B
Beta
0.019 0.003 1.362 0.000 1.2E-5 33.711
0.547 0.148 0.568 0.231 0.011 0.645 0.557
0.108 0.518 0.001 0.154 0.965
Winter noon
Coefficient
Sig.
B
Beta
0.011 0.023 0.278 0.000 0.001 30.014
0.153 0.519 0.054 0.192 0.291 0.867 0.834
0.450 0.001 0.556 0.059 0.061
Winter night
Coefficient
Sig.
B
Beta
0.013 0.004 2.393 7.1E-5 3.3E-5 1.831
0.372 0.197 0.762 0.084 0.029 0.303 0.128
0.433 0.571 0.027 0.737 0.931
Coefficient
Sig.
B
Beta
0.003 0.019 5.178 0.002 0.001 3.375
0.023 0.207 0.384 0.531 0.263
0.836 0.020 0.000 0.000 0.004
0.960 0.950
Abbreviations: BA ¼ percentage of building area, VC ¼ percentage vegetation cover, SVF ¼ sky view factor, DP ¼ distance to park, DW ¼ distance to water body, B ¼ unstandardized coefficient, Beta ¼ standardized coefficient, and Sig. ¼ significant level.
52
H. Yan et al. / Building and Environment 76 (2014) 44e53
temperature. The probable reasons are that during daytime, the air temperature variations at different locations were influenced by more factors, such as built-up ratio, percentage vegetation cover, and SVF, etc. At nighttime, however, the spatial pattern of temperature was simpler, and the air temperature difference may be mainly attributable to the cooling differences among different sites. The effects of land cover on ambient air temperatures have been well documented. Our results are consistent with those previous studies that the observed air temperature decreased as the percentage vegetation cover increased. However, the cooling effect of vegetation was stronger at night than during the day, although they varied slightly with the seasons. This is mainly because that the vegetation affects air temperature in different ways at different times. During the day, vegetation is one of the most influential factors in cooling temperature through partitioning solar radiation into latent heat rather than sensible heat. At this time, the shading of vegetation is also expected to produce a cooling effect on ambient air temperatures. During nighttime, however, the lower air temperature in vegetated area is mainly because of the greater radiative cooling rate. Thus, it seems that the cooling effect of vegetation produced by a higher cooling rate in the vegetated area at night could be larger than that produced by a combination of evapotranspiration and shading during the day [23,24]. In contrast to the cooling effect of vegetation, the increase of building area would significantly increase air temperature. The impact of building area is to favor partitioning of radiative energy into sensible rather than latent heat and to increase the importance of heat storage by the system, thus resulting in a higher temperature both during the day and night. In addition, an increase in building cover is also associated with the production of anthropogenic heat due to space heating, industrial operations or automobile use. Previous studies have also indicated that the intra-urban air temperature was related to urban geometry, as measured by H/W ratio or SVF. In the present study, the air temperature increased as the SVF increased during daytime, while a contrary tendency was observed during nighttime. We found no significant relationship between air temperature and SVF in both summer and winter daytime. This result diverge from the one found by Wong and Steve [43], which showed a significant and good correlation between air temperature and SVF during daytime. In our case, air measurements inside the park were carried out in open area with less or no shading from the trees, thus with higher SVF and with lower air temperature because of the evapotranspiration cooling from nearby vegetation, which can partly explain this difference in the results. During nighttime in winter, about 45% of the intra-urban variations in air temperature could be explained by the variation in SVF. However, SVF did not explain summer nighttime air temperature variations, most likely because the canopy of leaves in the summer has important effects on air temperature values in comparison to the values in winter. Some previous studies found that the correlation between SVF and air temperature is stronger in the defoliated season (i.e. winter season), because estimation of SVF is based only on building geometry and it show more connection with the real SVF situation in the winter than in the more vegetated summer season [30]. In summer, besides building geometry, street trees also reduce the openness to the sky. According to Wong and Steve [43], there is no adverse impact, i.e., reduction of nighttime net long-wave loss, due to the reduction of SVF by trees. It is possible that the tree shading result in smaller insolation because of altered shadow patterns during the daytime, which may influence the air temperature at night. It is usually expected that a site in green area is cooler than a site in the surrounding building areas. In our study, the park area was also found to be cooler than the surrounding environment both during daytime and nighttime in summer and winter. However, a
crucial issue to the value of green space, and in particular their impact on public health and energy saving, is whether a park has any effect on the temperature of the wider surrounding area [19]. Therefore, we analyzed the relationship between air temperature and distance to park and water body. We found that air temperature increased with increasing distance from the park boundary and more clearly in nighttime than in daytime. In Singapore, Chen and Wong [44] carried out measurements in two large park, and also found air temperature gradually increased with increasing distance from the park boundary. These results could be indicative of an extension of the park’s cooling effect into its surrounding, which would modify the urban thermal environment. The relationship between air temperature and distance to water was very similar to the relationship between air temperature and distance to park. A combination of two factors is likely to explain this relationship. First, a water body is moist and cool compared to its surrounding neighborhood and therefore may impact the microclimate surrounding it, which is quite similar to the park. Second, in the present study, the water body is located in the center of the park and its cooling effect would likely be affected by the park. The results of the multiple regression models indicate that it is the winter nighttime temperatures that can best be explained with the combination of percentage vegetation cover, SVF, distance to park and distance to water body. The explanatory power of the model is clearly weakest in the case of winter daytime temperature. In general, air temperature was more difficult to predict during the day than at night. Both in summer and winter daytime, the SVF was the only significant predictor of air temperature. The comparison of beta coefficients indicates that during this time, the SVF is also the most remarkable factor causing spatial differences in air temperatures. This result suggests that after controlling for all other factors, site geometry is one of the key variables in determining the urban thermal environment during daytime. During nighttime in summer, the percentage vegetation cover was the only significant predictor of air temperature, which means that the air temperature can be effectively mitigated by increasing the amount of vegetation cover. 5. Conclusions Field measurements were used to get the real temperature distribution across the study area during daytime and nighttime in summer and winter. Our results indicate spatial temperature difference between the maximum and minimum observed air temperature in the study area ranged from 1.2 to 7.0 C, depending on season and time of the day. The magnitude and spatial character of air temperatures, however, were mainly related to the microscale landscape design characteristics. Land cover composition was one of the key variables affecting local air temperature. The percentage of building area was consistently positively correlated with air temperature, whereas the percentage vegetation cover was negatively related. Site geometry was another crucial parameter because of its importance in determining the receipt and loss of radiation. The observed air temperature increased as SVF increased during daytime, while a contrary tendency was observed during nighttime. Our results also found that when testing the effects of the SVF on air temperature, it is important to control for the potential effects from other landscape variables, particularly the land cover features. In addition, the air temperature increased with increasing distance from the park and water body boundary. However, the results of the multivariate regression model indicated that the relative importance of these landscape parameters in explaining air temperature differences also varied among different times and seasons. These results have important implications for landscape design and planning, for both highly urbanized areas and areas
H. Yan et al. / Building and Environment 76 (2014) 44e53
where urbanization is still in process. It suggests that by designing landscapes that balance the relative amounts of building area and vegetation, and that optimize the canyon geometry and the spatial configuration of green space and water body, the impact of urbanization on microclimate can be greatly modified. Furthermore, in an era of global warming, these findings may provide important insights for planners and designers on mitigating the impact of urban development on urban heat island. Acknowledgments The authors would like to thank Aichun Song, Chenran Li, Zhuo Meng, Chaoqiong Wang and Li Feng for their help with data processing. The authors also thank Yuzhu Zeng and Lin Zhu for their assistance with the graphics. References [1] Kalnay E, Cai M. Impact of urbanization and land-use change on climate. Nature 2003;423(6939):528e31. [2] Chen XL, Zhao HM, Li PX, Yin ZY. Remote sensing image-based analysis of the relationship between urban heat island and land use/cover changes. Remote Sens Environ 2006;104(2):133e46. [3] Memon RA, Leung DYC, Chunho L. A review on the generation, determination and mitigation of urban heat island. J Environ Sci-China 2008;20:120e8. [4] Peng S, Piao S, Ciais P, Friedlingstein P, Ottle C, Breon FM, et al. Surface urban heat island across 419 global big cities. Environ Sci Technol 2011;46(2):696e 703. [5] Oke TR. The energetic basis of the urban heat island. Quart J R Met Soc 1982;108:1e24. [6] Saaroni H, Ben-Dor E, Bitan A, Potchter O. Spatial distribution and microscale characteristics of the urban heat island in Tel-Aviv, Israel. Landsc Urban Plan 2000;48(1):1e18. [7] Jonsson P. Vegetation as an urban climate control in the subtropical city of Gaborone, Botswana. Int J Climatol 2004;24(10):1307e22. [8] Jeong SY, Yoon SH. Method to quantify the effect of apartment housing design parameters on outdoor thermal comfort in summer. Build Environ 2012;53: 150e8. [9] Oke TR. Street design and urban canopy layer climate. Energy Build 1988;11(1):103e13. [10] Eliasson I. The use of climate knowledge in urban planning. Landsc Urban Plan 2000;48(1):31e44. [11] Brown RD. Ameliorating the effects of climate change: modifying microclimates through design. Landsc Urban Plan 2011;100(4):372e4. [12] Weng Q, Lu D, Schubring J. Estimation of land surface temperatureevegetation abundance relationship for urban heat island studies. Remote Sens Environ 2004;89(4):467e83. [13] Wong NH, Chen Y. Study of green areas and urban heat island in a tropical city. Habitat Int 2005;29(3):547e58. [14] Ng E, Chen L, Wang Y, Yuan C. A study on the cooling effects of greening in a high-density city: an experience from Hong Kong. Build Environ 2012;47: ́ 256e71. [15] Spronken-Smith RA, Oke TR. The thermal regime of urban parks in two cities with different summer climates. Int J Remote Sens 1998;19(11):2085e104. [16] Shashua-Bar L, Hoffman ME. Vegetation as a climatic component in the design of an urban street: an empirical model for predicting the cooling effect of urban green areas with trees. Energy Build 2000;31(3):221e35. [17] Chang CR, Li MH, Chang SD. A preliminary study on the local cool-island intensity of Taipei city parks. Landsc Urban Plan 2007;80(4):386e95.
53
[18] Wong NH, Jusuf SK, Win AL, Thu HK, Negara TS, Wu X. Environmental study of the impact of greenery in an institutional campus in the tropics. Build Environ 2007;42:2949e70. [19] Bowler DE, Buyung-Ali L, Knight TM, Pullin AS. Urban greening to cool towns and cities: a systematic review of the empirical evidence. Landsc Urban Plan 2010;97(3):147e55. [20] Oliveira S, Andrade H, Vaz T. The cooling effect of green spaces as a contribution to the mitigation of urban heat: a case study in Lisbon. Build Environ 2011;46(11):2186e94. [21] Cohen P, Potchter O, Matzarakis A. Daily and seasonal climatic conditions of green urban open spaces in the Mediterranean climate and their impact on human comfort. Build Environ 2012;51:285e95. [22] Krüger E, Givoni B. Outdoor measurements and temperature comparisons of seven monitoring stations: preliminary studies in Curitiba, Brazil. Build Environ 2007;42(4):1685e98. [23] Sun CY. A street thermal environment study in summer by the mobile transect technique. Theor Appl Climatol 2011;106(3e4):433e42. [24] Yokobori T, Ohta S. Effect of land cover on air temperatures involved in the development of an intra-urban heat island. Clim Res 2009;39(1):61e73. [25] Oke TR. Canyon geometry and the nocturnal urban heat island: comparison of scale model and field observations. J Climatol 1981;1(3):237e54. [26] Erell E, Williamson T. Intra-urban differences in canopy layer air temperature at a mid-latitude city. Int J Climatol 2007;27(9):1243e55. [27] Hart MA, Sailor DJ. Quantifying the influence of land-use and surface characteristics on spatial variability in the urban heat island. Theor Appl Climatol 2009;95(3e4):397e406. [28] Johnson DB. Urban modification of diurnal temperature cycles in Birmingham, UK. J Climatol 1985;5(2):221e5. [29] Yamashita S, Sekine K, Shoda M, Yamashita K, Hara Y. On relationships between heat island and sky view factor in the cities of Tama River basin, Japan. Atmos Environ 1986;20(4):681e6. [30] Unger J. Intra-urban relationship between surface geometry and urban heat island: review and new approach. Clim Res 2004;27(3):253e64. [31] Svensson MK. Sky view factor analysiseimplications for urban air temperature differences. Meteorol Appl 2004;11(03):201e11. [32] Eliasson I. Urban nocturnal temperatures, street geometry and land use. Atmos Environ 1996;30(3):379e92. [33] Upmanis H, Chen D. Influence of geographical factors and meteorological variables on nocturnal urban-park temperature differences-a case study of summer 1995 in Göteborg, Sweden. Clim Res 1999;13(2):125e39. [34] Upmanis H, Eliasson I, Lindqvist S. The influence of green areas on nocturnal temperatures in a high latitude city (Göteborg, Sweden). Int J Climatol 1998;18(6):681e700. [35] Nunez M, Oke TR. Energy balance of an urban Canyon. J Appl Meteorol 1977;16(1):11e9. [36] Taleghani M, Sailor DJ, Tenpierik M, van den Dobbelsteen A. Thermal assessment of heat mitigation strategies: the case of Portland State University, Oregon, USA. Build Environ 2014;73:138e50. [37] Ca VT, Asaeda T, Abu EM. Reductions in air conditioning energy caused by a nearby park. Energy Build 1998;29(1):83e92. [38] Jauregui E. Influence of a large urban park on temperature and convective precipitation in a tropical city. Energy Build 1991;15(3):457e63. [39] Saaroni H, Ziv B. The impact of a small lake on heat stress in a Mediterranean urban park: the case of Tel Aviv, Israel. Int J Biometeorol 2003;47(3):156e65. [40] Hamada S, Ohta T. Seasonal variations in the cooling effect of urban green areas on surrounding urban areas. Urban Urban Gree 2010;9(1):15e24. [41] Yu M, Liu Y, Dai Y, Yang A. Impact of urbanization on boundary layer structure in Beijing. Clim Change 2013;120:123e36. [42] Chow WTL, Roth M. Temporal dynamics of the urban heat island of Singapore. Int J Climatol 2006;26(15):2243e60. [43] Wong NH, Steve KJ. Air temperature distribution and the influence of sky view factor in a green Singapore estate. J Urban Plan D-ASCE 2009;136(3):261e72. [44] Chen Y, Wong NH. Thermal benefits of city parks. Energy Build 2006;38:105e 20.
Contents lists available at ScienceDirect
Building and Environment journal homepage: www.elsevier.com/locate/buildenv
Assessing the effects of landscape design parameters on intra-urban air temperature variability: The case of Beijing, China Hai Yan, Shuxin Fan, Chenxiao Guo, Fan Wu, Nan Zhang, Li Dong* College of Landscape Architecture, Beijing Forestry University, Beijing 100083, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 19 January 2014 Received in revised form 5 March 2014 Accepted 7 March 2014
Understanding the causes of the intra-urban air temperature variability is a first step in improving urban landscape design to ameliorate urban thermal environment. Here we investigated the spatial and temporal variations of air temperature at a local scale in Beijing, and their relationships with three categories of landscape design parameters, including the land cover features, site geometry, and spatial location. Air temperature measurements were conducted during the winter of 2012 and the summer of 2013 by mobile traverses. The results showed that spatial temperature difference between the maximum and minimum observed temperature in the study area ranged from 1.2 to 7.0 C, and varied depending on season and time of the day. The magnitude and spatial characteristic of the air temperature variations depend strongly on the landscape parameters characterizing the immediate environment of the measurement sites. Increasing the percentage vegetation cover could significantly decrease air temperature, while the increase of building area would significantly increase it. In addition, the observed air temperature increased as the sky view factor (SVF) increased during daytime, while a contrary tendency was observed during nighttime. However, the impacts of SVF on air temperature were context-dependent. Furthermore, the air temperature increased with increasing distance from the park and water body boundary. Our findings also indicated that the relative importance of these landscape parameters in explaining air temperature differences varied among different times and seasons. Therefore, if appropriately combined, all investigated landscape parameters can effectively improve urban thermal environment on a yearly basis. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Urban heat island Air temperature Landscape design Land cover Vegetation Sky view factor
1. Introduction Growing urban populations and urban expansion, with increasing built-up areas and human activities, results in significant modifications in the underlying surface properties and energy balance, and thus alters the urban climate [1]. A distinct feature of urban climate is the urban heat island (UHI) effect, which is directly related to the conversion of land cover from rural to urban covers [2e4]. Also, intra-urban surface cover variations result in the preferred spatial clustering of urban canopy layer (UCL) scale surfaces and surface units. This results in distinctive surface energy balance and microclimatic characteristics at a local scale [5]. Given the vast array of surface that forms the urban mosaic-like structure, it is not surprising that the microclimate within the urban environment varies significantly within the micro to local
* Corresponding author. Tel./fax: þ86 10 62336605. E-mail address: [email protected] (L. Dong). http://dx.doi.org/10.1016/j.buildenv.2014.03.007 0360-1323/Ó 2014 Elsevier Ltd. All rights reserved.
scale [6,7]. Urban microclimate both influences and is influenced by human behavior and decision-making, due to the complex interactions between air temperature and microscale landscape parameters [8]. Therefore, better understanding the causes of this intra-urban air temperature variability is critical in improving landscape design strategies to ameliorate urban thermal environment [9e11]. The first category of landscape parameters that influence urban microclimate is the land cover features. As urban areas develop and density increases, impervious urban surface covers replace permeable natural vegetation covers. Finally, impervious surface change the reflectivity and energy balance of surfaces resulting in locally higher air temperature. For this reason, vegetated areas play a critical role in moderating and dampening the warming effects of impervious land covers in urban environments [12e14]. Many field-based measurements have found that urban green areas are generally cooler than their surrounding built-up areas, and can produce air temperature differences up to 1e7 C, a phenomenon referred to as a “park cool island” [15e21]. Furthermore, a few
H. Yan et al. / Building and Environment 76 (2014) 44e53
qualitative associations between air temperature and its surrounding land cover features have been reported [22e24]. Sun [23], for example, showed that air temperature significantly correlated with green ratio and building ratio at night in Taichung City, Taiwan. Yokobori and Ohta [24] investigated the effect of land cover on ambient air temperatures and found that air temperatures varied significantly according to ambient land cover types, and air temperatures decreased as the amount of vegetated area around the sites increased. Site geometry is another crucial landscape parameter because of its importance in determining the receipt and loss of radiation. An important measure of surface geometry of a given site is the sky view factor (SVF), i.e. the fraction of the overlying hemisphere occupied by the sky [25]. During the day, the lower SVF at the canyon will result in less incoming solar radiation, and thus has an important effect on the ground surface temperature as well as the air temperature directly above it. The “urban cool island” phenomenon documented in many previous studies is mainly related to the low SVF [26,27]. During nighttime, the decreased SVF below roof level reduces the loss of long-wave radiation to the sky and also reduced turbulent heat transfer in the often calm canyon air. Therefore, theoretically it is considered to be a major factor of the nocturnal UHI phenomenon [28e31]. However, some studies have also found that SVF does not have a large importance on city temperature, especially not air temperature, and that it should not be assigned to much importance [32e34]. From past research it remains somewhat unclear the role that site geometry plays in its effects on urban air temperature. In addition, there can be no doubt about the important role of heat advection between contrasting surfaces at the microscale in the UCL [35]. Hence, the air temperature at a site is also affected by dispersion through turbulence and advection from its surroundings. In particular, vegetated areas or water bodies within the urban environment are moist and cool compared to their surrounding built areas [36], and thus the cooling effects of vegetation or water could extend into its surrounding area. For example, Ca et al. [37] investigated the influence of a park on the urban summer climate in the nearby areas, and found at noon, the park can reduce by up to 1.5 C the air temperature in a busy commercial area 1 km downwind. Observations made by other researches also show an extension of the cooling effect [16,34,38e40]. As this brief overview shows, the landscape design parameters of land cover, urban geometry, and the spatial location (proximity to the park or water) have been found to influence the local air temperature. The problem, however, is that different researchers look at the problem form a different angle using different landscape parameters, and it is very difficult to conclude which particular landscape factors would be more important in effecting the air temperature within the urban context. In addition, urban climate is also affected by external influences such as the topographic features, season and prevailing weather. Therefore, it is important to control geographical, seasonal and meteorological (i.e. wind speed and cloud cover) variables as much as possible to understand the location specific impacts on changes in urban air temperature. In this study, we investigate the spatial and temporal differences of air temperature at a local scale in Beijing, and to gain insights on its linkage to the landscape parameters including percentage of building area, percentage vegetation cover, SVF, distance to park, and distance to water body. Specifically, this study tried to answer the following two questions: (1) how do the different landscape parameters differentially contribute to explaining the variance of local air temperatures, and (2) if the relative importance of these landscape parameters in explaining local air temperature varies by season and time of day?
45
2. Methodology 2.1. Study area and measurement sites Beijing (39 560 N, 116 200 E), the capital of China, is located in the northern part of the North China Plain. It is the second largest city in China with a total population of 19.6 million by the end of 2010. It has a monsoon influenced humid continental climate characterized by hot and humid summers and generally cold, windy and dry winters. According to the climatological normals (1971e2000), January is the coldest month with an average temperature of 3.7 C, while July is the hottest month with an average temperature of 26.2 C. The highest temperature in 1971e2000 was 41.9 C, and the coldest temperature was 18.3 C. The main wind direction is from southeast to northeast in summer and in reverse during winter. Since 1978, Beijing’s urban population and yearly construction area have been gradually increasing. This urbanization process, with its increasing built-up areas and anthropogenic activities, results in significant modifications in the underlying surface properties and the quick increase in the intensification of the UHI effect [41]. The measurements took place at the Beijing Olympic Park and its surrounding built-up areas in the Beijing city (Fig. 1). The study area is rectangular, 3.2 2.1 km in size and situated at the northern part of the city. It can be regarded as a typical area in the process of urbanization in Beijing. This area is very flat, and thus temperature difference due to topographic influence can be ignored. In addition, this district has a perfectly orthogonal urban geometry and the streets are oriented EeW and NeS, with an average width of 25 m. Therefore, it is a good study area to explore the relationship between air temperature and landscape design factors. In this study area, 26 measurement sites were chosen with the intention to investigate the quantitative relationship between air temperature and landscape variables (Fig. 1). A description of the measurement sites is given in Table 1. These measurement sites located sufficiently close to one another to be affected by uniform meso-scale climate conditions, and yet, also affected by distinctly different micro-scale landscape characteristics. The nature of the adjacent underlying surface greatly affects air temperature. Thus, to standardize temperature, all air temperature measurements were made over a common surface: paved road, as shown in Fig. 2(a). 2.2. Air temperature measurements Mobile traverses were used to conduct air temperature measurements on 8 days during the winter of 2012 (December 2012 to February 2013) and the summer of 2013 (June to August). Survey times were early afternoon (14:00e15:00) and midnight (23:00e 00:00). All observations were carried out during clear and windless days, i.e. when the UHI effects could be expected to be most pronounced. Avoidance of windy and cloudy conditions also minimizes the influence of meteorological variables. Measurements of air temperature were taken using a mobile station with a thermistor temperature sensor connected to a data logger with a precision of 0.1 C along the measurement route. The temperature sensor was fitted within a radiation shield, which mounted onto the front of a bicycle at 1.5 m above ground level. A GPS travel recorder was used to note geographical coordinates of latitude, longitude, and altitude synchronized to observations. Three fixed weather stations were installed in different regions within the study area for recording of the daily trend of air temperature. These data were used to correct the values taken with the mobile station because the measurements at different sites were not instantaneous. Each traverse took about 1 h to complete, and all data were adjusted to the beginning time of each transect by
46
H. Yan et al. / Building and Environment 76 (2014) 44e53
Fig. 1. Location of the study area and measurement sites. The numbers represent the mobile measurement sites, while the lowercase letters represent the three fixed air temperature measurement sites.
modifying values using temperature change rates over the study area based on fixed weather station measurements of air temperature. Wind speed and direction were obtained from the weather observatory station near the measurement sites. In most cases, the wind does not change significantly during the individual one-hour measurement session. 2.3. Measurement and computations of landscape parameters Numerous parameters have been developed to assess and quantify the effects of urban environment characteristics on air temperature. For this study, three categories of landscape parameters including land cover feature, site geometry and spatial location, were selected to measure site environmental characteristics. These parameters were selected based on the following principles: (1) the potential effects on air temperature, (2) easily calculated, (3) readily understood, easily controlled by design professionals, and (4) minimal redundancy. Land cover features included the percentage of building area and percentage vegetation cover. Site geometry was measured using SVF and spatial location was measured using distance to park and distance to water body. Aerial photographs (Google Maps, September 15, 2012) were used to acquire the land cover features in the study area. Since the
variation in air temperature with regard to the land cover composition of each measurement sites need to be quantified, a buffer zone was used to accomplish this. According to the results of some previous studies, air temperature in an urban location is mainly affected by its surrounding surface cover with a few hundred meters in radius, and the far field tends to have negligible impacts due to the intricate urban fabrics [22,24]. Thus, a buffer zone with 150 m in radius was used in our study. Using a similar method to Krüger and Givoni [22], the percentage of building area and vegetation cover for each measurement site were calculated with AutoCAD, by drawing corresponding areas on the aerial photograph (Fig. 2(b)). For the determination of the SVF, fisheye images at each measurement site were taken by a fisheye lens (Sigma 8 mm circular fisheye lens) coupled to a Cannon EOS 5D Mark II digital camera. The fisheye images were processed into black and white image, in which the sky is white and the building and trees are black. Then, the images were put into RayMan software for the SVF calculation. The SVF were calculated separately for summer and winter because of the seasonal variation in vegetation, as shown in Fig. 2(c and d). The effect of park and water body has been estimated based on the straight distance between each measurement site and the edge of the park and the water body. In the present study, the park refers mainly to the Beijing Olympic Forestry Park, and the water body refers mainly to the Dragon-shaped Olympic Water System in
H. Yan et al. / Building and Environment 76 (2014) 44e53
47
Table 1 Landscape characteristics and description of measurement sites. Site number
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Landscape parameters
Description of measurement sites
BA (%)
VC (%)
SVF (summer)
SVF (winter)
DP (m)
DW (m)
23.4 28.6 29.7 32.1 31.9 7.3 7.1 0.0 4.2 9.5 13.2 0.6 11.2 17.8 28.9 28.7 31.4 24.7 12.3 21.6 10.3 32.1 32.8 12.4 1.3 10.3
33.1 31.3 27.8 27.4 25.4 53.0 57.6 82.1 62.2 54.7 20.1 44.8 65.9 20.1 27.5 15.7 39.7 34.6 36.7 16.6 27.9 19.8 30.2 42.7 71.8 61.0
0.63 0.62 0.59 0.66 0.49 0.70 0.49 0.66 0.72 0.76 0.39 0.76 0.58 0.61 0.59 0.32 0.31 0.11 0.59 0.46 0.76 0.51 0.38 0.38 0.46 0.46
0.67 0.71 0.60 0.67 0.60 0.80 0.75 0.82 0.79 0.79 0.64 0.83 0.77 0.60 0.60 0.47 0.51 0.55 0.75 0.46 0.76 0.50 0.45 0.49 0.75 0.60
180 660 1259 1375 1084 612 128 5 180 626 1019 1259 1004 613 505 625 1006 1242 1340 918 907 828 204 44 5 45
702 1097 1202 949 606 619 458 321 292 128 18 212 396 410 695 1004 1010 639 282 919 90 722 870 452 35 437
Broad street, multi-storey buildings Broad street, different-sized buildings High-rise and multi-storey buildings Broad street, multi-storey buildings Broad street, high-rise buildings Park’s perimeter road, tree covered Park’s perimeter road, tree covered Inside park, open grass area Inside park, open paved area Inside park, open paved area Waterside, tree covered Open area, parking lot Broad street, street trees Broad street, high-rise buildings High-rise buildings, tree covered High-rise buildings, tree covered High-rise buildings, tree covered Multi-storey buildings, tree covered Broad street, high-rise buildings High-rise building, heavy traffic Open area High-rise buildings, heavy traffic High-rise buildings, tree covered High-rise buildings, park edge Inside park, open vegetated area Multi-story building, park edge
Note: BA, VC, SVF, DP, and DW refer to the percentage of building area, percentage vegetation cover, sky view factor, distance to park, and distance to water body, respectively.
Beijing City. The detailed statistics of these landscape parameters are listed in Table 1. 2.4. Statistical analyses We first performed simple linear regression to analyze the bivariate relationship between air temperatures and percentage of building area, percentage vegetation cover, SVF, distance to park and distance to water body. Further, a multivariate regression analysis was used to quantify the relative contribution of five landscape variables to the air temperature differences. The predictive model was assumed as:
Y ¼ a þ b1 BA þ b2 VA þ b3 SVF þ b4 DP þ b5 DW
(1)
Where Y represents the value of air temperature ( C); a is a constant; b1, b2.b5 are the coefficients for each independent variable; BA (%) is the percentage of building area; VA (%) is the percentage vegetation area; SVF is the sky view factor; DP (m) is the distance to park; and DW (m) is the distance to water body. All of the statistical analyses were performed using SPSS 17.0 software. 3. Results 3.1. Temporal and spatial variation of air temperatures The spatial patterns of the average temperature of four mobile traverses among different times and seasons within the study area are presented in Fig. 3. The air temperature differences among different locations are very significant. During daytime in summer (Fig. 3(a)), the hottest place was at site 4 with mean air temperature reaching 35.5 C. This site is located in a shallow street canyon, which is fully exposed to solar radiation during daytime. The continuous heating of the pavement surface by solar radiation and the release of anthropogenic heat contributes to the high air
temperature here. It is worth noting that the air temperatures measured at sites 16 and 17, which are located in the center of the commercial area, were not much higher even with the expected high anthropogenic heat flux. This is because the high-rise buildings located on both sides of the street and mature street trees can provide very good shading. At this time, the coolest sites were observed at sites 8 and 25, both of which were located inside the park with dense vegetation cover. Even though these sites were exposed to solar radiation during the daytime, evapotranspiration from the grass surface and trees caused low air temperature. During nighttime in summer (Fig. 3(b)), the pattern of temperature distribution in the study area is very clear. The park area now became a cool island surrounded by hot areas. The mean air temperature difference between the lowest inside the park and the highest in the building area was 3.2 C. During daytime in winter, the air temperatures distribution profile is rather complicated, as shown in Fig. 3(c). At this time, sites 1, 14 and 19 were found to display a higher air temperature than other sites. The probable reasons are that they are exposed to the open space, a reduction of building shade at their immediate environment and with increased exposure to direct sunshine. The air temperature at site 24 was the lowest as it is located at the south-side of the EeW oriented street canyon, well shaded by highrise apartment buildings. During night in winter (Fig. 3(d)), the hottest site was site 22, which is located in a heavy traffic street characterized by low ventilation, high business activities, multistoried buildings and heavy road traffic. The coldest site was observed at the north-central part of the park, where mean air temperature reached 8.5 C. This open space stores less heat during daytime and releases the heat during nighttime at a faster rate compared to densely built environment. To analyze the temporal variation of temperature, DTmax-min, which is the difference between the maximum and minimum observed temperature, was calculated for each measurement. Fig. 4 clearly shows that the differences of the DTmax-min among different
48
H. Yan et al. / Building and Environment 76 (2014) 44e53
Fig. 2. Landscape characteristics of one example of the measurement sites (site 1). (a) measurement environment, (b) land cover features, (c) SVF in summer, and (d) SVF in winter.
times and seasons are very significant. In summer, the daytime DTmax-min ranged from 1.7 to 1.9 C, and it ranged from 2.1 to 4.5 C during nighttime. In winter, the daytime DTmax-min ranged from 1.2 to 2.1 C, and the nighttime DTmax-min ranged from 4.3 to 7.0 C. On average, the DTmax-min was largest in winter nighttime (5.9 C), followed by summer nighttime (3.4 C), summer daytime (1.8 C), and winter daytime (1.7 C). To explore whether the driving mechanisms of air temperature differences varied with times and seasons, we further analyze the relationship of air temperature between daytime and nighttime, and summer and winter. From the aspect of diurnal pattern of spatial difference in air temperature, nighttime air temperature increased significantly with daytime air temperature in summer
(Fig. 5(a); R2 ¼ 0.45, P < 0.001). However, there was no relationship between daytime and nighttime air temperature in winter (Fig. 5(b); R2 ¼ 0.00, P ¼ 0.990). For seasonal pattern of spatial difference in air temperature, there was no relationship between summer and winter daytime air temperature (Fig. 5(c); R2 ¼ 0.08, P ¼ 0.154), while there was a significant positive relationship between summer and winter nighttime air temperature (Fig. 5(d); R2 ¼ 0.75, P < 0.001). 3.2. Influence of the land cover features on air temperature The effects of the composition of land cover features on air temperatures vary clearly among different times and seasons, as
Fig. 3. Spatial variation of average air temperatures (average of four measurement days) by observation site at (a) summer daytime, (b) summer nighttime, (c) winter daytime, and (d) winter nighttime.
H. Yan et al. / Building and Environment 76 (2014) 44e53
49
building area (Fig. 6(c); R2 ¼ 0.664, P < 0.001) and a strong negative relationship between air temperature and the percentage vegetation cover (Fig. 6(d); R2 ¼ 0.620, P < 0.001). 3.3. Influence of the site geometry on air temperature
Fig. 4. Box plots showing differences between the warmest and coldest observed temperature among different times and seasons. The bulging dots in the middle indicate the median, the boxes indicate the upper and lower quantiles, and the whiskers indicate the range.
shown in Fig. 6. During daytime in summer, there was a weak positive relationship between air temperature and the percentage of building area (Fig. 6(a); R2 ¼ 0.256, P ¼ 0.008) and a weak relationship between air temperature and the percentage vegetation cover (Fig. 6(b); R2 ¼ 0.202, P ¼ 0.021). During nighttime in summer, there was a strong positive relationship between air temperature and the percentage of building area (Fig. 6(a); R2 ¼ 0.674, P < 0.001) and a strong negative relationship between air temperature and the percentage vegetation cover (Fig. 6(b); R2 ¼ 0.702, P < 0.001). During daytime in winter, there was no significant relationship between air temperature and both of percentage of building area and percentage vegetation cover (Fig. 6(c and d)). During nighttime in winter, there was a strong positive relationship between air temperature and the percentage of
During daytime in both summer and winter, air temperatures increased with increasing SVF (Fig. 7), suggesting that during daytime, when there was solar radiation, the lower the sky openness, the lower the air temperature was. However, the simple linear regression analysis did not show a statistical significant relationship between air temperature and SVF in both summer daytime (Fig. 7(a); R2 ¼ 0.100, P ¼ 0.115) and winter daytime (Fig. 7(b); R2 ¼ 0.002, P ¼ 0.811). During nighttime, it is clearly shown that the correlation has changed into negative, which means that during this time, the net outgoing long-wave decreases at location that has a low SVF values, resulting in higher temperatures. According to the obtained statistical measurements, about 45.2% (P < 0.001) of the intra-urban variations in air temperature can be explained by the variation in SVF during winter nighttime (Fig. 7(b)). SVF, however, does not explain summer nighttime air temperature variations (Fig. 7(a); R2 ¼ 0.048, P ¼ 0.281). 3.4. Influence of the distance to park and water on air temperature Fig. 8 shows the relationships between air temperature and distance to park and distance to water body. During daytime and nighttime in summer and winter, air temperature increased with increasing distance to park, but more rapidly at nighttime than at daytime and more rapidly at winter nighttime than at summer nighttime. The R2 were 0.248, 0.327, and 0.486 during summer daytime, summer nighttime, and winter nighttime, respectively, indicating that the variable (distance to park) account for 24.8%, 32.7%, and 48.6% of variance in the air temperature distribution. However, there was no significant relationship between air temperature and distance to park during daytime in winter. The air temperatures determined by the variable of distance to water was
Fig. 5. Relationships between (a) summer daytime temperature and summer nighttime temperature, (b) winter daytime temperature and winter nighttime temperature, (c) summer daytime temperature and winter daytime temperature, and (d) summer nighttime temperature and winter nighttime temperature.
50
H. Yan et al. / Building and Environment 76 (2014) 44e53
Fig. 6. Relationships between air temperature and land cover features for the percentage of building area and percentage vegetation cover in summer (a, b) and in winter (c, d).
very similar to that determined by the distance to park. The R2 were 0.149, 0.471, and 0.462 during summer daytime, summer nighttime, and winter nighttime, respectively, and there was also no significant relationship between air temperature and distance to water at winter daytime. 3.5. Multiple regression analysis Multiple regression analysis was carried out in order to find out how well the observed air temperature differences could be explained by the combination of the five landscape variables (Table 2). The regression results offer insights into the influence of different variables on air temperature among different times and seasons. The coefficient of determination (R2) represented the proportion of the variation in air temperature could be explained by the regression models; and the standardized coefficients (Beta coefficients) of predictive models represented the relative contributions of different landscape variables to the air temperature difference. During daytime in summer, approximately 64.5% of the variation in air temperature was explained jointly by the five landscape variables. The only significant factor in this model was SVF. The comparison of beta coefficients indicates that the SVF was also the most important variable. According to this model, for every 10% increase in SVF will increase air temperature by 0.14 C. During nighttime in summer, about 86.7% of the variation in air temperature was explained jointly by the five landscape variables. The only significant factor in this model was percentage vegetation cover. A 10% increase in vegetation cover decreased air temperature by 0.23 C. During daytime in winter, approximately 30.3% of the variation in air temperature was explained jointly by the five landscape variables. The only significant factor in this model was SVF. The comparison of beta coefficients indicates that the SVF was also the most important variable. According to this model, for every 10% increase in SVF will increase air temperature by 0.24 C. During nighttime in summer, about 86.7% of the variation in air temperature was explained jointly by the five landscape variables. The significant independent variables were the percentage
vegetation cover, SVF, distance to park and distance to water. The negative coefficients of percentage vegetation cover and SVF suggest that an increase in the percentage vegetation cover and SVF will decrease air temperature. In contrast, the positive coefficients of distance to park and water indicate that temperature will increase with the increase of distance from park and water body boundary. 4. Discussion Results from the surveys demonstrate large spatial differences in temperatures in the study area. The magnitude and spatial characteristic of the differences varied depending on season and time of the day. At nighttime, the pattern of temperature distribution in the study area was very clear. The park area became a cool island surrounded by hot built areas. However, the spatial pattern by day tended to be less well defined. It is also worth noting that at this time, air temperature at high-rise building areas might be cooler than the sites inside the park, resulting in an urban cool island. This phenomenon has also been observed by Chow and Roth [42], mainly because of the slower warming of urban surfaces due to the solar heat storage of building materials and the shading effect of the building and trees. The maximum air temperature differences between the warmest and coldest site were largest in winter nighttime, followed by summer nighttime, summer daytime, and winter daytime. These basic patterns are generally similar to many previous urban heat island and park cool island studies conducted in cities of various sizes. Air temperature differences were greater at night, especially in winter during clear, calm weather conditions. This is largely attributable to the difference in radiative cooling rates between natural vegetation and built area. The vegetated area was more exposed to the sky than the building area, which allows for a higher cooling rate of this area. At the same time, the decrease in temperature in the building area was lower because this area is surrounded by buildings, which influences the loss of long-wave radiative to the sky. Since the difference in radiative cooling rates between the vegetated area and the building area was particularly pronounced in winter, when the canopies are leafless and at the
H. Yan et al. / Building and Environment 76 (2014) 44e53
51
Fig. 7. Relationships between air temperature and SVF in summer (a) and in winter (b).
Fig. 8. Relationships between air temperature and distance to park and distance to water body in summer (a, b) and in winter (c, d).
same time as lower soil moisture leads to low thermal heat capacity in the vegetated area, a maximum temperature difference of 7.0 C occurred during nighttime in winter. We further analyzed the correlation of mean air temperature pattern between different times and seasons. There was a significant relationship between nighttime air temperature and daytime air temperature in summer, which suggests that there is inertia of
daytime processed into the following night. However, we found no relationship between daytime and nighttime air temperature in winter, which means that the factors driving air temperature variations during the day are different than those at night in winter. We also found that there was no relationship between summer and winter daytime air temperature, while there was a significant positive relationship between summer and winter nighttime air
Table 2 Regression results of air temperature and five landscape variables. The bold figures are significant variable with p < 0.05. Variables
Summer noon
Summer night
Coefficient
BA VC SVF DP DW Constant R2 Adjusted R2
Sig.
B
Beta
0.019 0.003 1.362 0.000 1.2E-5 33.711
0.547 0.148 0.568 0.231 0.011 0.645 0.557
0.108 0.518 0.001 0.154 0.965
Winter noon
Coefficient
Sig.
B
Beta
0.011 0.023 0.278 0.000 0.001 30.014
0.153 0.519 0.054 0.192 0.291 0.867 0.834
0.450 0.001 0.556 0.059 0.061
Winter night
Coefficient
Sig.
B
Beta
0.013 0.004 2.393 7.1E-5 3.3E-5 1.831
0.372 0.197 0.762 0.084 0.029 0.303 0.128
0.433 0.571 0.027 0.737 0.931
Coefficient
Sig.
B
Beta
0.003 0.019 5.178 0.002 0.001 3.375
0.023 0.207 0.384 0.531 0.263
0.836 0.020 0.000 0.000 0.004
0.960 0.950
Abbreviations: BA ¼ percentage of building area, VC ¼ percentage vegetation cover, SVF ¼ sky view factor, DP ¼ distance to park, DW ¼ distance to water body, B ¼ unstandardized coefficient, Beta ¼ standardized coefficient, and Sig. ¼ significant level.
52
H. Yan et al. / Building and Environment 76 (2014) 44e53
temperature. The probable reasons are that during daytime, the air temperature variations at different locations were influenced by more factors, such as built-up ratio, percentage vegetation cover, and SVF, etc. At nighttime, however, the spatial pattern of temperature was simpler, and the air temperature difference may be mainly attributable to the cooling differences among different sites. The effects of land cover on ambient air temperatures have been well documented. Our results are consistent with those previous studies that the observed air temperature decreased as the percentage vegetation cover increased. However, the cooling effect of vegetation was stronger at night than during the day, although they varied slightly with the seasons. This is mainly because that the vegetation affects air temperature in different ways at different times. During the day, vegetation is one of the most influential factors in cooling temperature through partitioning solar radiation into latent heat rather than sensible heat. At this time, the shading of vegetation is also expected to produce a cooling effect on ambient air temperatures. During nighttime, however, the lower air temperature in vegetated area is mainly because of the greater radiative cooling rate. Thus, it seems that the cooling effect of vegetation produced by a higher cooling rate in the vegetated area at night could be larger than that produced by a combination of evapotranspiration and shading during the day [23,24]. In contrast to the cooling effect of vegetation, the increase of building area would significantly increase air temperature. The impact of building area is to favor partitioning of radiative energy into sensible rather than latent heat and to increase the importance of heat storage by the system, thus resulting in a higher temperature both during the day and night. In addition, an increase in building cover is also associated with the production of anthropogenic heat due to space heating, industrial operations or automobile use. Previous studies have also indicated that the intra-urban air temperature was related to urban geometry, as measured by H/W ratio or SVF. In the present study, the air temperature increased as the SVF increased during daytime, while a contrary tendency was observed during nighttime. We found no significant relationship between air temperature and SVF in both summer and winter daytime. This result diverge from the one found by Wong and Steve [43], which showed a significant and good correlation between air temperature and SVF during daytime. In our case, air measurements inside the park were carried out in open area with less or no shading from the trees, thus with higher SVF and with lower air temperature because of the evapotranspiration cooling from nearby vegetation, which can partly explain this difference in the results. During nighttime in winter, about 45% of the intra-urban variations in air temperature could be explained by the variation in SVF. However, SVF did not explain summer nighttime air temperature variations, most likely because the canopy of leaves in the summer has important effects on air temperature values in comparison to the values in winter. Some previous studies found that the correlation between SVF and air temperature is stronger in the defoliated season (i.e. winter season), because estimation of SVF is based only on building geometry and it show more connection with the real SVF situation in the winter than in the more vegetated summer season [30]. In summer, besides building geometry, street trees also reduce the openness to the sky. According to Wong and Steve [43], there is no adverse impact, i.e., reduction of nighttime net long-wave loss, due to the reduction of SVF by trees. It is possible that the tree shading result in smaller insolation because of altered shadow patterns during the daytime, which may influence the air temperature at night. It is usually expected that a site in green area is cooler than a site in the surrounding building areas. In our study, the park area was also found to be cooler than the surrounding environment both during daytime and nighttime in summer and winter. However, a
crucial issue to the value of green space, and in particular their impact on public health and energy saving, is whether a park has any effect on the temperature of the wider surrounding area [19]. Therefore, we analyzed the relationship between air temperature and distance to park and water body. We found that air temperature increased with increasing distance from the park boundary and more clearly in nighttime than in daytime. In Singapore, Chen and Wong [44] carried out measurements in two large park, and also found air temperature gradually increased with increasing distance from the park boundary. These results could be indicative of an extension of the park’s cooling effect into its surrounding, which would modify the urban thermal environment. The relationship between air temperature and distance to water was very similar to the relationship between air temperature and distance to park. A combination of two factors is likely to explain this relationship. First, a water body is moist and cool compared to its surrounding neighborhood and therefore may impact the microclimate surrounding it, which is quite similar to the park. Second, in the present study, the water body is located in the center of the park and its cooling effect would likely be affected by the park. The results of the multiple regression models indicate that it is the winter nighttime temperatures that can best be explained with the combination of percentage vegetation cover, SVF, distance to park and distance to water body. The explanatory power of the model is clearly weakest in the case of winter daytime temperature. In general, air temperature was more difficult to predict during the day than at night. Both in summer and winter daytime, the SVF was the only significant predictor of air temperature. The comparison of beta coefficients indicates that during this time, the SVF is also the most remarkable factor causing spatial differences in air temperatures. This result suggests that after controlling for all other factors, site geometry is one of the key variables in determining the urban thermal environment during daytime. During nighttime in summer, the percentage vegetation cover was the only significant predictor of air temperature, which means that the air temperature can be effectively mitigated by increasing the amount of vegetation cover. 5. Conclusions Field measurements were used to get the real temperature distribution across the study area during daytime and nighttime in summer and winter. Our results indicate spatial temperature difference between the maximum and minimum observed air temperature in the study area ranged from 1.2 to 7.0 C, depending on season and time of the day. The magnitude and spatial character of air temperatures, however, were mainly related to the microscale landscape design characteristics. Land cover composition was one of the key variables affecting local air temperature. The percentage of building area was consistently positively correlated with air temperature, whereas the percentage vegetation cover was negatively related. Site geometry was another crucial parameter because of its importance in determining the receipt and loss of radiation. The observed air temperature increased as SVF increased during daytime, while a contrary tendency was observed during nighttime. Our results also found that when testing the effects of the SVF on air temperature, it is important to control for the potential effects from other landscape variables, particularly the land cover features. In addition, the air temperature increased with increasing distance from the park and water body boundary. However, the results of the multivariate regression model indicated that the relative importance of these landscape parameters in explaining air temperature differences also varied among different times and seasons. These results have important implications for landscape design and planning, for both highly urbanized areas and areas
H. Yan et al. / Building and Environment 76 (2014) 44e53
where urbanization is still in process. It suggests that by designing landscapes that balance the relative amounts of building area and vegetation, and that optimize the canyon geometry and the spatial configuration of green space and water body, the impact of urbanization on microclimate can be greatly modified. Furthermore, in an era of global warming, these findings may provide important insights for planners and designers on mitigating the impact of urban development on urban heat island. Acknowledgments The authors would like to thank Aichun Song, Chenran Li, Zhuo Meng, Chaoqiong Wang and Li Feng for their help with data processing. The authors also thank Yuzhu Zeng and Lin Zhu for their assistance with the graphics. References [1] Kalnay E, Cai M. Impact of urbanization and land-use change on climate. Nature 2003;423(6939):528e31. [2] Chen XL, Zhao HM, Li PX, Yin ZY. Remote sensing image-based analysis of the relationship between urban heat island and land use/cover changes. Remote Sens Environ 2006;104(2):133e46. [3] Memon RA, Leung DYC, Chunho L. A review on the generation, determination and mitigation of urban heat island. J Environ Sci-China 2008;20:120e8. [4] Peng S, Piao S, Ciais P, Friedlingstein P, Ottle C, Breon FM, et al. Surface urban heat island across 419 global big cities. Environ Sci Technol 2011;46(2):696e 703. [5] Oke TR. The energetic basis of the urban heat island. Quart J R Met Soc 1982;108:1e24. [6] Saaroni H, Ben-Dor E, Bitan A, Potchter O. Spatial distribution and microscale characteristics of the urban heat island in Tel-Aviv, Israel. Landsc Urban Plan 2000;48(1):1e18. [7] Jonsson P. Vegetation as an urban climate control in the subtropical city of Gaborone, Botswana. Int J Climatol 2004;24(10):1307e22. [8] Jeong SY, Yoon SH. Method to quantify the effect of apartment housing design parameters on outdoor thermal comfort in summer. Build Environ 2012;53: 150e8. [9] Oke TR. Street design and urban canopy layer climate. Energy Build 1988;11(1):103e13. [10] Eliasson I. The use of climate knowledge in urban planning. Landsc Urban Plan 2000;48(1):31e44. [11] Brown RD. Ameliorating the effects of climate change: modifying microclimates through design. Landsc Urban Plan 2011;100(4):372e4. [12] Weng Q, Lu D, Schubring J. Estimation of land surface temperatureevegetation abundance relationship for urban heat island studies. Remote Sens Environ 2004;89(4):467e83. [13] Wong NH, Chen Y. Study of green areas and urban heat island in a tropical city. Habitat Int 2005;29(3):547e58. [14] Ng E, Chen L, Wang Y, Yuan C. A study on the cooling effects of greening in a high-density city: an experience from Hong Kong. Build Environ 2012;47: ́ 256e71. [15] Spronken-Smith RA, Oke TR. The thermal regime of urban parks in two cities with different summer climates. Int J Remote Sens 1998;19(11):2085e104. [16] Shashua-Bar L, Hoffman ME. Vegetation as a climatic component in the design of an urban street: an empirical model for predicting the cooling effect of urban green areas with trees. Energy Build 2000;31(3):221e35. [17] Chang CR, Li MH, Chang SD. A preliminary study on the local cool-island intensity of Taipei city parks. Landsc Urban Plan 2007;80(4):386e95.
53
[18] Wong NH, Jusuf SK, Win AL, Thu HK, Negara TS, Wu X. Environmental study of the impact of greenery in an institutional campus in the tropics. Build Environ 2007;42:2949e70. [19] Bowler DE, Buyung-Ali L, Knight TM, Pullin AS. Urban greening to cool towns and cities: a systematic review of the empirical evidence. Landsc Urban Plan 2010;97(3):147e55. [20] Oliveira S, Andrade H, Vaz T. The cooling effect of green spaces as a contribution to the mitigation of urban heat: a case study in Lisbon. Build Environ 2011;46(11):2186e94. [21] Cohen P, Potchter O, Matzarakis A. Daily and seasonal climatic conditions of green urban open spaces in the Mediterranean climate and their impact on human comfort. Build Environ 2012;51:285e95. [22] Krüger E, Givoni B. Outdoor measurements and temperature comparisons of seven monitoring stations: preliminary studies in Curitiba, Brazil. Build Environ 2007;42(4):1685e98. [23] Sun CY. A street thermal environment study in summer by the mobile transect technique. Theor Appl Climatol 2011;106(3e4):433e42. [24] Yokobori T, Ohta S. Effect of land cover on air temperatures involved in the development of an intra-urban heat island. Clim Res 2009;39(1):61e73. [25] Oke TR. Canyon geometry and the nocturnal urban heat island: comparison of scale model and field observations. J Climatol 1981;1(3):237e54. [26] Erell E, Williamson T. Intra-urban differences in canopy layer air temperature at a mid-latitude city. Int J Climatol 2007;27(9):1243e55. [27] Hart MA, Sailor DJ. Quantifying the influence of land-use and surface characteristics on spatial variability in the urban heat island. Theor Appl Climatol 2009;95(3e4):397e406. [28] Johnson DB. Urban modification of diurnal temperature cycles in Birmingham, UK. J Climatol 1985;5(2):221e5. [29] Yamashita S, Sekine K, Shoda M, Yamashita K, Hara Y. On relationships between heat island and sky view factor in the cities of Tama River basin, Japan. Atmos Environ 1986;20(4):681e6. [30] Unger J. Intra-urban relationship between surface geometry and urban heat island: review and new approach. Clim Res 2004;27(3):253e64. [31] Svensson MK. Sky view factor analysiseimplications for urban air temperature differences. Meteorol Appl 2004;11(03):201e11. [32] Eliasson I. Urban nocturnal temperatures, street geometry and land use. Atmos Environ 1996;30(3):379e92. [33] Upmanis H, Chen D. Influence of geographical factors and meteorological variables on nocturnal urban-park temperature differences-a case study of summer 1995 in Göteborg, Sweden. Clim Res 1999;13(2):125e39. [34] Upmanis H, Eliasson I, Lindqvist S. The influence of green areas on nocturnal temperatures in a high latitude city (Göteborg, Sweden). Int J Climatol 1998;18(6):681e700. [35] Nunez M, Oke TR. Energy balance of an urban Canyon. J Appl Meteorol 1977;16(1):11e9. [36] Taleghani M, Sailor DJ, Tenpierik M, van den Dobbelsteen A. Thermal assessment of heat mitigation strategies: the case of Portland State University, Oregon, USA. Build Environ 2014;73:138e50. [37] Ca VT, Asaeda T, Abu EM. Reductions in air conditioning energy caused by a nearby park. Energy Build 1998;29(1):83e92. [38] Jauregui E. Influence of a large urban park on temperature and convective precipitation in a tropical city. Energy Build 1991;15(3):457e63. [39] Saaroni H, Ziv B. The impact of a small lake on heat stress in a Mediterranean urban park: the case of Tel Aviv, Israel. Int J Biometeorol 2003;47(3):156e65. [40] Hamada S, Ohta T. Seasonal variations in the cooling effect of urban green areas on surrounding urban areas. Urban Urban Gree 2010;9(1):15e24. [41] Yu M, Liu Y, Dai Y, Yang A. Impact of urbanization on boundary layer structure in Beijing. Clim Change 2013;120:123e36. [42] Chow WTL, Roth M. Temporal dynamics of the urban heat island of Singapore. Int J Climatol 2006;26(15):2243e60. [43] Wong NH, Steve KJ. Air temperature distribution and the influence of sky view factor in a green Singapore estate. J Urban Plan D-ASCE 2009;136(3):261e72. [44] Chen Y, Wong NH. Thermal benefits of city parks. Energy Build 2006;38:105e 20.