Urban vegetation for reducing heat related mortality.

Environmental Pollution xxx (2014) 1e10

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Urban vegetation for reducing heat related mortality Dong Chen a, *, Xiaoming Wang a, Marcus Thatcher b, Guy Barnett a, Anthony Kachenko c, Robert Prince c a

CSIRO Climate Adaptation Flagship and CSIRO Ecosystem Sciences, Melbourne, Australia CSIRO Climate Adaptation Flagship and CSIRO Marine and Atmospheric Research, Melbourne, Australia c Nursery & Garden Industry Australia, Sydney, Australia b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 January 2014 Received in revised form 1 May 2014 Accepted 3 May 2014 Available online xxx

The potential benefit of urban vegetation in reducing heat related mortality in the city of Melbourne, Australia is investigated using a two-scale modelling approach. A meso-scale urban climate model was used to quantify the effects of ten urban vegetation schemes on the current climate in 2009 and future climates in 2030 and 2050. The indoor thermal performance of five residential buildings was then simulated using a building simulation tool with the local meso-climates associated with various urban vegetation schemes. Simulation results suggest that average seasonal summer temperatures can be reduced in the range of around 0.5 and 2 C if the city were replaced by vegetated suburbs and parklands, respectively. With the limited buildings and local meso-climates investigated in this study, around 5e28% and 37e99% reduction in heat related mortality rate have been estimated by doubling the city’s vegetation coverage and transforming the city into parklands respectively. Crown Copyright Ó 2014 Published by Elsevier Ltd. All rights reserved.

Keywords: Climate change Mortality rate Urban vegetation Urban heat island Heat waves

1. Introduction Heat waves have been recognized as one of the major natural hazards and kill more people than any other natural hazards in Australia (PwC, 2011; State of Australian Cities (2013)). The heat wave event in Melbourne, Australia during the summer of 2009 may have resulted in hundreds of excess deaths over what would normally be expected for the period (DHS, 2009). In 2003, the European heatwave resulted in over 70,000 excess deaths (Robine et al., 2008). Indeed, the linkage between mortality and heat has been long recognized (Changnon et al., 1996; Haines et al., 2006; Luber, 2008; Huang et al., 2012). Several earlier studies have tried to quantify the relationship between climate conditions and mortality rate in Australia (Nicholls et al., 2008; Tong et al., 2010; Loughnan et al., 2010; Tong et al., 2014). Nicholls et al. (2008) analysed the mortality rate in Melbourne for people over 65 from 1979 to 2001. They reported that excess heat related mortality amongst the population over 65 years of age may increase rapidly when the mean daily temperatures (the average of yesterday’s maximum and this morning’s minimum) exceeded about 30 C.

* Corresponding author. E-mail address: [email protected] (D. Chen).

The impact of heat waves in major Australian cities such as Melbourne and Sydney is likely to become worse due to further urbanization in existing suburbs, the effects of global warming and the ageing population. Further urbanization may potentially intensify the urban heat island (UHI) effect, a phenomenon whereby urbanized population centers experience warmer temperatures compared with surrounding rural areas (Morris et al., 2001; Coutts et al., 2007). At the same time, global warming projections suggest the likely increase in the number of warm nights, heat wave frequency and duration in Australia (CSIRO, 2007; Alexander and Arblaster, 2008). Consequently, it is essential to develop effective strategies to protect urban communities from the impact of heat waves. In recent years, urban greening and cool surfaces have attracted substantial interests and show promise in mitigating the impact of heat waves (Rosenfeld et al., 1998; Akbari et al., 2001; Liu and Bass, 2005; Rosenzweig et al., 2006; Yu and Hien, 2006; Luber, 2008; Alexandri and Jones, 2008; Memon et al., 2008; Bowler et al., 2010; Wong and Lau, 2013). Urban greening can mitigate heat wave impacts by trees that shade buildings and cool the ambient air by evapotranspiration. Cool surfaces such as reflective roofs and paving surfaces can further reduce heat absorption in the urban area. Urban greening and cool surfaces can also reduce the cooling energy demand and thus green house gas (GHG) emissions in cities (Ca et al., 1998; Susca et al., 2011; Xu et al., 2012; Smith et al., 2012).

http://dx.doi.org/10.1016/j.envpol.2014.05.002 0269-7491/Crown Copyright Ó 2014 Published by Elsevier Ltd. All rights reserved.

Please cite this article in press as: Chen, D., et al., Urban vegetation for reducing heat related mortality, Environmental Pollution (2014), http:// dx.doi.org/10.1016/j.envpol.2014.05.002

2

D. Chen et al. / Environmental Pollution xxx (2014) 1e10

Previous researches in heat related mortality have been focused on the relationship between the ambient weather conditions and the mortality rate (Nicholls et al., 2008; Tong et al., 2010, 2014). Using urban climate modelling, Kalkstein et al. (2014) investigated the cooling effects due to different urban vegetation schemes and surface reflectivity. Based on the relationship between the ambient weather conditions and heat related mortality rate, they estimated that 10% increase in the urban vegetation coverage and surface reflectivity can result in an average 7% reduction in mortality during heat waves in the District of Columbia. It should be noted that heat related death occurs both indoors and outdoors. Cadot et al. (2007) reported that 74% of excess deaths during the 2003 summer heat wave in Paris occurred among those who were living at home. They emphasized that people most at risk from dying were those aged over 75 years and living alone. Although there is no available information on the locations of heat related excess deaths in Australia, the situation in Australia may be similar to that in Paris considering that the most vulnerable population is the elderly people group (DHS, 2009). Due to the importance of thermal comfort and health safety of the living and working environment, there have been numerous studies on thermal comfort and heat stress in buildings since early last century (Brager and de Dear, 1998; Epstein and Moran, 2006; ASHRAE, 2009; Djongyang et al., 2010). Based on human body heat balance calculations and empirical observations, around 40 indices have been developed for defining thermal comfort and heat stress (Epstein and Moran, 2006). Generally, using these heat stress indices, thermal comfort and heat stress can be qualitatively categorized into different levels. For example, levels of heat stress may be categorized into four groups using the Discomfort Index (DI) which is defined as the mean of the air dry bulb and wet bulb temperatures, i.e., DI < 22, no heat stress; 22 DI < 24, mild sensation of heat; 24 DI < 28, hot with difficulties in physical work; DI 28, severe heat with risk for heat illness (Epstein and Moran, 2006). Qualitatively, an indoor environment with high DI index, especially for DI over 28, may potentially present high risks in causing heat illness including heat related mortality. However, quantitative relationships between heat related mortality rate and the DI index, or any other heat stress index, in residential buildings are not available. At least two reasons may attribute to the difficulties in obtaining such relationships. First, heat stress indices are normally developed based on the heat balance for maintaining the body-core temperature around 37 C or empirical tests with relatively healthy population groups such as workers in the working environment (Epstein and Moran, 2006). However, the sensations and responses of vulnerable population groups to heat stress are different from healthy population groups (Hwang et al., 2007). Second, information about the indoor thermal environment in which heat related deaths occurred is hardly available. Consequently, although urban greening may have the potential in mitigating the impact of heat waves, there is currently no established methodology in quantifying its benefit in reducing heat related mortality in individual buildings. In this study, the role of urban vegetation in reducing heat related mortality in the city of Melbourne, Australia was investigated using a two-scale modelling approach. First, a meso-scale urban climate model was used for quantifying the effects of ten different urban vegetation schemes on the local climate in Melbourne. Then, the indoor thermal performance of five individual residential buildings was simulated using a building simulation tool using these vegetation modified local climates. The potential reduction in heat related mortality rate was then estimated from the correlation between the simulated indoor mean daily temperature and 20 year daily mortality rate records.

2. Methodologies 2.1. Urban climate with various urban vegetation schemes In this study, the long term average impact on the urban climate was first established with urban climate modelling for the Melbourne Central Business District (CBD) area assuming various urban vegetation schemes. Hourly climate files for 2009, 2030 and 2050 with various urban vegetation schemes were then generated for building thermal performance simulations using a ‘morphing’ approach. 2.1.1. Urban climate model A recently developed urban climate model (UCM-TAPM) (Thatcher and Hurley, 2012) was used for the investigation of the impact of urban vegetation schemes on local urban climate. The UCM-TAPM combines an urban canopy model (UCM) with The Air Pollution Model (TAPM), a meso-scale climate model developed by CSIRO (Hurley et al., 2005). The UCM includes an efficient big-leaf model to represent in-canyon vegetation in the predominately suburban component of Australian cities. The model employs a multiple one-way nesting procedure to dynamically downscale meteorological reanalyses/forecasts, typically in steps of 30 km, 10 km, 3 km and 1 km. The meteorological component of UCM-TAPM is nested within synoptic-scale analyses/ forecasts which drive the model at the boundaries of the outer grids. In the UCM-TAPM, a grid tile of the land surface can assign one of 39 surface types that include a wide range of natural and built surface types such as water body, forest, shrub land, grassland, pasture, CBD, urban, and industrial. The characteristics of the surface types such as the average building height, building height to street canopy width ratio, vegetation coverage, leaf area index, surface albedos etc. can be adjusted for specific urban surface conditions. The UCM-TAPM has been validated against the measurements from several urban and rural weather stations (Thatcher and Hurley, 2012). Using UCM-TAPM, the impact of various urban vegetation schemes on the local urban climate can be quantified. When modelling the ‘current’ Melbourne climate under various urban forms and vegetation schemes, the downscaled reanalysis climate data from the National Centre for Environmental Prediction (NCEP) were used for the lateral boundaries of the outer grids for the UCM-TAPM. For ‘future’ Melbourne climates under various urban forms and vegetation schemes, UCM-TAPM simulations were carried out with the boundary conditions based on the downscaled climate data from GFDL2.1 under the A2 emission scenario. 2.1.2. Urban vegetation schemes As illustrated by the green square in Fig. 1a, the Melbourne CBD area is represented by nine 1 km 1 km grids in the UCM-TAPM. Fig. 1b shows the corresponding map for Melbourne which covers an area approximately 12.5 12.5 km2 surrounding the Melbourne CBD area. In this study, four level nesting grids were used in the UCM-TAPM model, i.e., 30 km 30 km, 10 km 10 km, 3 km 3 km and 1 km 1 km grids. There are 25 25 grids for each grid level and the total simulation domain covers an area of 750 750 km2 with a spatial resolution of 1 km at the finest grid level. The impact of urban vegetation on the Melbourne CBD local climate was investigated using the UCM-TAPM by replacing the Melbourne CBD areas with the ten urban forms listed in Table 1. The CBD urban form (Urban Form Number 6) represents the existing Melbourne CBD with the vegetation and building coverage percentages estimated from Google images. Urban Form Number 9, i.e., CBD with 50% green roof, assumes 50% green roof coverage on all the building roofs in the Melbourne CBD area. Although



3

Fig. 1. Representation of the Melbourne CBD: (a) image from the UCM-TAPM; (b) Map of Melbourne.

redeveloping the Melbourne CBD area into some of the urban forms listed in Table 1 would be unrealistic, it is important to understand the potential effects of various urban vegetation schemes for assisting effective and resilient future urban designs. 2.1.3. Climate file preparation Modelling building indoor thermal performance with building simulation tools required weather files for the local climate in consideration. Typical Meteorological Year (TMY) weather files are widely adopted as the reference weather files for assessing building thermal performance. TMY weather data are specially selected so that it represents the range of weather phenomena for a given location. TMY weather files for the Melbourne CBD area were derived with the weather data from or near the Melbourne

Regional Office of the Australian Bureau of Meteorology for the period from 1976 to 2004 which is approximately centered at around 1990 (ABCB, 2006). Future weather files for 2030 and 2050 were prepared using the ‘morphing’ approach developed by Belcher et al. (2005). With the morphing approach, weather data in the TMY weather files were modified to take into account the climate change effect using the flowing equations:

T ¼ T0 þ DTm þ aTm ðT0 hT0 im Þ

(1)

RH ¼ RH0 þ DRHm

(2)

I ¼ 1 þ am;I I0

(3)

Table 1 The urban forms and vegetation schemes investigated in this study. Urban form number

Urban form

Vegetation type

Vegetation coverage (%)

Leaf area index

Building coverage (%)

Average building height (m)

Building height to canyon width Ratiod

Irrigatione

1 2 3 4 5 6 7 8 9

Forest (low sparse) Shrub-land Grassland Urban (leafy) Urban (generic) CBD CBD (with 1/3 Vegetation) CBD (Double Vegetation) CBD (50% Green Roof)

Low sparse (woodland) Mid dense (scrub) Mid dense tussock Mixed Mixed Mixed Mixed Mixed Mixed

100 100 100 49 38a 15b 5 33 15

0 0 0 40 45a 65b 65 62 65

e e e 6.0 6.0 12.0 12.0 12.0 12.0

e e e 0.4 0.4 1.3 1.3 1.3 1.3

No No No Yes Yes Yes Yes Yes Yes

10

CBD (Double Vegetation þ 50% Green Roof)

Mixed

2.0 2.6 1.2 3 3 3 3 3 3 1.5c 3 1.5c

62

12.0

1.3

Yes

a b c d e

33

From Coutts et al. (2007). Estimated from Google images. Areas other than vegetation and buildings are accounted as roads in the UCM-TAPM. Leaf area index for green roof vegetation. Estimated average building height divided by the street/road width. Rainfall has been modelled and considered in the UCM-TAPM modelling. Irrigation is the additional watering to prevent vegetation dry out.


4


V ¼ 1 þ am;v V0

further improved if granular mortality data for the city of Melbourne is available in the future.

(4)

where T, RH, I and V are the projected future hourly dry-bulb ambient air temperature ( C), relative humidity (%), solar irradiance (W/m2) and wind speed (m/s) respectively; D represents changes in the corresponding weather parameters projected by the UCM-TAPM with GFDL2.1 under the A2 emission scenario; am,I and am,v are the monthly averaged percentage changes in solar irradiance and wind speed (%), again projected by the UCM-TAPM with GFDL2.1 under the A2 emission scenario; subscript 0 for the

2.2.2. Building thermal performance In this study, the thermal performance of five residential buildings was modelled by AccuRate e a residential house energy rating software used in Australia (Delsante, 2005). AccuRate was developed by coupling a frequency response building thermal model (Walsh and Delsante, 1983) and a multi-zone ventilation model (Ren and Chen, 2010) for thermal performance and energy requirement calculation of residential buildings. The five residential buildings are assumed to operate without heating and air conditioning. It is also assumed that the occupants operate the windows and doors based on the following rules:

m DTMINm reference climate and m for monthly-mean; aTm ¼ DhTTMAX ; 0max i hT0min i m

m

hT0 im ; hT0max im ; hT0min im are the monthly-mean values of the ambient dry-bulb temperature, daily maximum temperature and daily minimum temperature from the TMY weather data, respectively. DTm, DTMAXm, DTMINm are the projected changes in the monthly-mean values of ambient dry-bulb temperature, daily maximum temperature and daily minimum temperature ( C) respectively due to climate change. When the impact of urban vegetation schemes are considered, DTm, DTMAXm and DTMINm are obtained by the sum of the corresponding temperature change due to projected climate change and that due to different vegetation schemes as listed in Table 1. In order to investigate the impact of various vegetation schemes during the 2009 summer in Melbourne, the weather file for the CBD urban form from 1 July 2008 to 30 June 2009 was constructed using the weather data obtained from the Melbourne Regional Office of the Bureau of Meteorology. This weather file was then modified using the DTm, DTMAXm and DTMINm obtained for different urban vegetation schemes by the UCM-TAPM simulations which will be further discussed in Section 3.1.

Windows and doors are closed if indoor air temperature is below 22 C; If indoor air temperature is above 24 C and outdoor air temperature is below indoor air temperature, windows and doors are opened. Otherwise, windows and doors are closed.

2.2. Indoor thermal environment and mortality

In Melbourne, detached houses represent around 76% of the residential housing stock and the remaining 24% consists of buildings such as semi-detached buildings, flats, units, apartments (ABS, 2011). In this study, five residential buildings were used which include three detached houses, a semi-detached three bedroom two-storey townhouse and a two bedroom apartment at the top of a two-storey building. It is assumed that there was no insulation in these buildings to represent the low-end housing stock in Melbourne with their occupants potentially exposed to high heat stress risks during heat weave periods. Table 2 describes the design specifications of the five buildings.

2.2.1. Mortality and population data Mortality data from 1988 to 2007 for Melbourne were obtained from the Health and Vitals Statistics Unit, Australian Bureau of Statistics (ABS). Due to privacy protection, the data obtained from ABS were limited for the Melbourne Statistical Division (SD) of usual residence by sex and by two age groups, being 0e75 and 75þ. Melbourne SD covers the metropolitan area of Melbourne as well as its surrounding urban fringe and rural areas, which including the Dandenong Ranges, the Yarra Valley and the Mornington Peninsula. It has a population exceeding 3.5 million and accounts for over 70% of the population of the state of Victoria. In order to calculate the mortality rate, the corresponding populations by sex by the two age groups from 1988 to 2007 in Melbourne SD were also obtained from ABS. In this study, constrained by the availability of mortality data, the average mortality rate obtained for Melbourne SD is taken to be the mortality rate for the city of Melbourne. It is understood that this assumption may reduce the accuracy of the current study. It is believed that using the same methodology adopted in this study, the accuracy of the findings presented in this research can be

2.2.3. Correlate indoor mean air temperature with mortality rate In order to understand the linkage between indoor air temperature and mortality rate in Melbourne, hourly simulations were carried out for 20 years from 1 January 1988 to 31 December 2007 for the five buildings facing north, east, south and west. The hourly weather data used for the 20 year simulations were based on the records from the Bureau of Meteorology. Considering that occupants are normally in the living room during daytime and in the bedroom at nighttime, the mean daily temperature, Tm, daily, for a building is defined as the average of yesterday’s daytime (after 7am) maximum in the living room and this morning’s (before 7am) minimum in the master bedroom. In the 20 years from 1 January 1988 to 31 December 2007, there are a total of 7305 mean daily temperatures for each building grouped into consecutive 0.5 C bands. The average mortality rates for the city of Melbourne corresponding to a particular mean daily temperature band were obtained from the mortality rate of the days when the mean daily temperatures sit in a particular 0.5 C mean daily temperature band. For example, the average mortality rate

Table 2 Brief descriptions of the five buildings investigated in this study. Number of bedroom

Floor area (m2)

External wall

Internal walls

House1 House2 House3 Townhouse

4 3 3 3

161 88 88 108

Plasterboard Plasterboard Plasterboard Plasterboard

Apartment

2

64

Brick veneer Brick Veneer Brick Veneer Brick veneer and weather board Brick veneer

on on on on

studs studs studs studs

Plasterboard on studs

Floor

Ceiling

Windows

Concrete slab Concrete slab Timber floor to subfloor Concrete slab

Plasterboard Plasterboard Plasterboard Plasterboard

Timber, single-glazed Aluminium, single-glazed Aluminium, single-glazed Aluminium, single-glazed

Timber floor to ground floor underneath

Plasterboard

Aluminium, single-glazed



5

corresponding to the mean daily temperature band from 28 C to 28.5 C is the average of the mortality rates for the city of Melbourne for all the days (in the 20 years) when the building has the mean daily temperature between 28 C and 28.5 C. The relationships between the mean daily temperatures and the average mortality rate can be established from the 20 year daily mortality rate and the simulated daily mean temperature for the five buildings in four directions for both males and females. This approach for correlating mortality rate with the indoor mean daily temperature is similar to those used by several researchers (Donaldson et al., 2003; McMichael et al., 2006; Tong et al., 2014) except that the correlations established in previous studies were between mortality rate and the ambient air mean daily temperature. These correlations between the indoor mean daily temperatures and the average mortality rate were used to estimate the mortality rate in 2009, 2030 and 2050 in Section 3.2. 3. Results and discussions 3.1. The impact of urban vegetation on local climate UCM-TAPM simulations were carried out for the Melbourne CBD area in 2001, 2009, 2030, 2047, 2050, 2087 and 2090 with the ten urban forms as described in Table 1. It is noted that like other urban climate models, the UCM-TAPM is used for predicting urban climate probability distributions and seasonal variability of the climatic variables, especially when UCM-TAPM is nested in Global Climate Models predicting global warming scenarios rather than reanalyses of observed weather conditions (Thatcher and Hurley, 2012). The UCM-TAPM is not intended for predicting short term climate phenomena such as the hourly air temperature or hourly wind speed. In this study, the impact of a specific urban form on the local climate in the Melbourne CBD area is expressed as the changes in three summer seasonal air temperatures obtained by the UCMTAPM simulations for this particular urban form relative to that for the CBD urban form (Urban Form Number 6), i.e., DTMean, DTMin and DTMax. Here, TMin: the average summer daily minimum temperature (average over January, February and December); TMean: the average summer mean temperature; and TMax: the average summer daily maximum temperature. Fig. 2aec shows the DTMin, DTMean and DTMax obtained by the UCM-TAPM simulations for various urban forms in ‘current’ years in 2001 and 2009 using the NECP reanalysis climate data and for five future years in 2030, 2047, 2050, 2087 and 2090 projected using GFDL2.1. In Fig. 2, DV and GR are the abbreviations for double vegetation and green roof respectively. The thick green lines are the mean values of the DTMin, DTMean and DTMax for each urban form across the seven years. It is seen that the average summer daily maximum temperature in the Melbourne CBD area reduces with the increase in urban vegetation coverage with the generic urban area predicted to be around 0.5 C cooler than the existing CBD urban forms. The maximum cooling potential of around 2 C could be achieved if the CBD area would have been transformed to a natural forest park land. Similar cooling potential for urban vegetation was observed for the corresponding TMin and TMean in the CBD area. It was found that despite the projected warming trends in the future, for a given urban vegetation scheme, the variations in DTMin, DTMean and DTMax are generally within 20% of the means for the seven years. The cooling effect of urban vegetation is determined by two aspects, vegetation shading and

Fig. 2. Projections of the DTMin, DTMean and DTMax for the ten urban forms in 2001, 2009, 2030, 2047, 2050, 2087 and 2090: (a) DTMin ( C); (b) DTMean ( C); (c) DTMax ( C).


6


evapotranspiration, with the latter the dominant effect since the vegetation shading effect for a specific urban form is more or less fixed. Evapotranspiration rate of urban vegetation scenarios can vary due to variation in the rainfall in each year. However, as shown in Fig. 2, overall, these variations in vegetation shading and evapotranspiration have limited effect on the DTMin, DTMean and DTMax for a specific vegetation scheme. Consequently, as mentioned in Section 2, the weather files for a particular year with a specific urban vegetation scheme can be constructed using the ‘morphing’ approach with the DTm, DTMAXm and DTMINm obtained by the UCM-TAPM. In this study, the DTm, DTMAXm and DTMINm for a specific urban vegetation scheme were the average values obtained from the seven year UCM-TAPM simulation results. With this approach, the constructed weather file retains the long term average temperature changes due to a specific urban vegetation scheme and projected climate change.

MRexcess;House1;north;veg

scheme

¼

X Tm;daily >28:5 C

h MRTm;daily ;House1;north;veg

3.2. Excess mortality rate due to indoor heat stress As discussed in Section 2.2.3, 20 relationships between the mean daily temperatures and the average mortality rate can be established from 20 year daily mortality rate and the simulated daily mean temperature for the five building in four directions for both males and females. Fig. 3 shows the relationships between the mean daily temperature, Tm,daily, and the average mortality rates for males and females over 75 years old for House1 facing north and south respectively. Similar to the “U” shape relationship observed between the mortality rate and the mean daily ambient air temperatures by many previous studies (Donaldson et al., 2003; McMichael et al., 2006; Tong et al., 2014), extremely cold and hot indoor mean daily temperatures in buildings also correspond to high average mortality rates. Similar trends were observed for the remaining buildings with different facing directions. For each set of the relationship between the average mortality rates and the mean daily indoor temperatures, a correlation was obtained using the R Software (version 3.0.2) as shown in Fig. 3. Consequently, a total of 20 correlations were obtained for males and females over 75 years old respectively for these five buildings in four directions. These

MRimpact;House1veg

scheme

¼

MRexcess;House1;veg

scheme

MRexcess;

MRexcess;House1;CBD

correlations are used to estimate the mortality rate in 2009, 2030 and 2050 in the following. It is noted that the correlations obtained as shown in Fig. 3 are the long term average relationships between the mean daily temperature, Tm,daily, and the average mortality rates and may not be applicable to a specific heat wave event. It is also noted that each plot in Fig. 3 was obtained by assuming that the Melbourne CBD area has only one type of building which faces a single direction. This assumption does not adequately reflect the present urban environment in Melbourne. Consequently, Fig. 3 should be

interpreted with care as it does not necessary suggest that a particular mean daily temperature in a building with a particular facing direction will cause the corresponding mortality rate in this building. Nevertheless, Fig. 3 suggests that the high mortality rate during a heat wave period may correlate with the high mean daily temperatures in various individual buildings. From Fig. 3 for House1 and results for other four buildings, it was found that the average mortality rate increases rapidly when the mean daily indoor temperature Tm,daily is approximately 28.5 C. Therefore, in this study, the mortality rate at Tm,daily ¼ 28.5 C is assumed as the base mortality rate. The excess mortality rate due to heat stress was then estimated as the sum of the mortality rate for those days with mean daily temperature of the building above 28.5 C minus the base mortality rate. For example, the excess mortality rate for the north facing House1 for a specific urban vegetation form is calculated as:

scheme

MRTm;daily ¼28:5 o C;House1;north;veg

scheme

DaysTm;daily

i

(5)

Here, MRexcess; House1; north;veg scheme is the excess heat related mortality rate for a north facing House1 for a specific urban vegetation scheme. DaysTm;daily is the number of days when the mean daily temperature Tm,daily is in a 0.5 C temperature band above 28.5 C. The estimated excess mortality rate for each building type is obtained as the average of the four facing directions of this building type, e.g., the excess mortality rate for House1 is calculated as:

MRexcess;House1;veg

scheme

¼

1 MRexcess;House1;north;veg scheme 4 þ MRexcess;House1;east;veg scheme þ MRexcess;House1;south;veg þ MRexcess;House1;west;veg

scheme

scheme

(6) The potential impact on excess heat related mortality rate for a specific urban vegetation scheme was then calculated as the percentage difference of the excess mortality rate in comparison with that for the CBD urban scheme as follows:

House1;CBD

100%

(7)

This potential impact on excess mortality rate may be considered as an estimation of the difference in the heat related mortality rate when the Melbourne CBD area has a specific urban vegetation scheme relative to the CBD scheme. Fig. 4 show the potential impact on excess mortality rate for males and females over 75 respectively with different urban vegetation schemes in 2009, 2030 and 2050 for each of the five buildings. Although there are variations among different buildings across different years, the overall trends are consistent in that urban vegetation can potentially reduce the excess mortality rate due



7

Fig. 3. Relationships between the mean daily temperature in House1 and the average mortality rate in Melbourne from 1 January 1988 to 31 December 2007: (a) North facing, males; (b) North facing, females; (c) South facing, males; (d) South facing, females.

to heat stress. In general, the reduction in excess mortality rate increases with an increase in vegetation coverage and intensity with the forest scheme (assuming the whole Melbourne CBD area has a forest scheme) achieving the best performance from 37% up to 99% reduction in the excess mortality rate. Results for the five buildings show that increasing the Melbourne CBD vegetation coverage from 15% to 33% may reduce the average heat related mortality rate in the range between 5% and

28%. It is understood that there is no similar research in Melbourne for the benefit of green coverage on heat related mortality. However, this predicted reduction in the average heat related mortality rate is comparable with the results presented in Kalkstein et al. (2014), who estimated that 10% increase in the urban vegetation coverage and surface reflectivity can result in an average 7% reduction in heat related mortality during heat waves in the District of Columbia.


D. Chen et al. / Environmental Pollution xxx (2014) 1e10 30%

30%

10%

10%

-10% 2009 House1 2009 House2 2009 House3 2009 TownHouse 2009 Apartment

-30% -50% -70% -90%

ReducƟon in mortality Rate (%)


8

-110%

-10%

-50% -70% -90% -110%

(b) 30%

10%

10%

-10% -30%

-70%

2030 House1 2030 House2 2030 House3 2030 TownHouse 2030 Apartment

-90%



(a) 30%

-50%


-30%

-10% -30% 2030 House1 2030 House2 2030 House3 2030 TownHouse 2030 Apartment

-50% -70% -90%

-110%

-110%

(d) 30%

10%

10%

-10% -30% -50% -70%


-90% -110%



(c) 30%

-10% -30% -50% -70%


-90% -110%

(e)

(f)

Fig. 4. Estimation of the reduction in heat related mortality rate with various urban vegetation schemes for: (a) males in 2009; (b) females in 2009; (c) males in 2030; (d) females in 2030; (e) males in 2050; (f) females in 2050.



It was also found that the reduction in the mortality rate for males is only slightly different from that for females for a given urban vegetation scheme in a specific building. This is due to two reasons. First, in a given building, both the male and female occupants are assumed to expose to the same mean daily temperature. Second, from Eqs. (1)e(3), it can be shown that for a specific house in a given urban vegetation scheme, if the relationship between mortality rate and the mean daily temperature is exactly linear with a gradient of k, the potential impact on excess heat related mortality rate obtained by Eq. (3) is only dependent on the number of days with Tm,daily > 28.5 C for this urban vegetation scheme and for the CBD scheme. The potential impact on excess heat related mortality rate obtained by Eq. (3) does not change with the gradient k. As shown in Fig. 3, the relationship between mortality rate and the mean daily temperature is essentially linear for both males and females after Tm,daily ¼ 28.5 C. Consequently, the percentage reductions in the mortality rate for a given building are approximately the same for males and females for a given urban vegetation scheme in a specific year in comparison with the CBD vegetation scheme. However, the percentage reductions in the mortality rate for both male and female vary among different buildings or the same building in different years, since different buildings and different years (thus different weather conditions) result in different number of days with Tm,daily > 28.5 C for a given urban vegetation scheme and for the CBD vegetation scheme. This finding suggests that urban greening may have similar benefit in reducing heat related mortality rate for both males and females. Considering that it is a first attempt in quantifying urban vegetation in mitigating heat related mortality rate at the scale of buildings, the current study may be improved in several aspects in the future. First, there are uncertainties involved in future climate projections and thus the future weather files used for building thermal performance in this study. This may be improved in the future by using multiple Atmosphere-Ocean General Circulation Models (AOGCMs) with different carbon emissions scenarios to take into account the uncertainties of individual AOGCM models as recommended by the Intergovernmental Panel on Climate Change (IPCC, 2007). Second, in this study, the average mortality rate obtained for Melbourne SD is taken to be the mortality rate for the city of Melbourne. Using the same methodology, the results can be improved if granular mortality data for the city of Melbourne can be available in the future. Third, uncertainty analysis in the results may be carried out by considering the confidence level in the correlations between the mean daily temperature, Tm,daily, and the average mortality rates. Fourth, as discussed in Section 3.1, the building sample size in this study is small and the findings from this study will need to be further verified with large building sample sizes. Nevertheless, the research findings are encouraging and show the potential benefit of urban vegetation in reducing the heat related excess mortality rate by mitigating the impact of heat waves and projected global warming. 4. Conclusions The potential for urban vegetation in reducing heat related mortality for 2009 and the projected future climates in 2030 and 2050 in the city of Melbourne, Australia was investigated for ten different urban vegetation schemes using a two-scale modelling approach: a meso-scale urban climate model and a building thermal performance model. Simulation results showed that the average seasonal summer temperatures can be reduced in the range of around 0.5 and 2 C if Melbourne CBD were replaced by vegetated suburbs and planted parklands, respectively. It was also found that despite the projected warming in the future, the average

9

seasonal cooling potential due to various urban vegetation schemes remains reasonably unchanged in comparison with those predicted for the current climate, indicating little dependency on climate change. Simulations of the indoor thermal environment for five residential buildings in 2009 and for projected future climates in 2030 and 2050 showed that urban vegetation can potentially reduce excess mortality rate in the city of Melbourne due to heat stress. An increase in the Melbourne CBD vegetation coverage from 15% to 33% may reduce the average heat related mortality rate in the range between 5% and 28%. Reduction in the excess mortality rate in the range from 37% up to 99% is estimated by replacing the whole Melbourne CBD area with forest parkland. Results also suggest that urban greening may have similar benefit in the percentage reduction in heat related mortality rate for both males and females. Although variations were predicted among different buildings and in different years, these findings may provide confidence as well as a methodology for urban planners in mitigating urban heat island and heat wave impact with urban greening strategies. The modelling approach used in this study is a first attempt in quantifying urban vegetation in mitigating heat related mortality rate at the scale of buildings. As discussed in Section 3.2, the current methodology can be further improved in at least the following four aspects in the future: take into account the uncertainties involved in future climate projections by using multiple AOGCMs with different carbon emissions scenarios; take into account the confidence level of correlation between the mean daily temperature and the average mortality rates; use granular mortality data for the city of Melbourne for establishing the correlation between the mortality rate and indoor thermal environment; and, extend the building sample size. Further studies using the two-scale modelling approach is also required to validate our results. Acknowledgement This study was partly funded by the Horticulture Australia Limited using the Nursery Industry Levy (Project # NY11013 and NY12018) and CSIRO Climate Adaptation Flagship. The authors would like to thank Mr Zhiwei Xu in Queensland University of Technology on using R software and Mr Felix Lipkin in CSIRO for providing the map of Melbourne. References ABCB, 2006. Protocol for House Energy Rating Software. Australian Building Codes Board. Akbari, H., Pomerantz, M., Taha, H., 2001. Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas. Sol. Energy 70, 295e310. Alexander, L.V., Arblaster, J., 2008. Assessing trends in observed and modelled climate extremes over Australia in relation to future projections. Int. J. Climatol. 29 (3), 417e435. Alexandri, E., Jones, P., 2008. Temperature decreases in an urban canyon due to green walls and green roofs in diverse climates. Build. Environ. 43, 480e493. ASHRAE Handbook, 2009. 2009 Fundamentals. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc. Australia, 2011. Census, Australian Bureau of Statistics (ABS). http://www.abs.gov.au (accessed in January 2014). Belcher, S.E., Hacker, J.N., Powell, D.S., 2005. Constructing design weather data for future climates. Build. Serv. Eng. Res. Technol. 26 (1), 49e61. Bowler, D.E., Buyung-Ali, L., Knight, T.M., Pullin, A.S., 2010. Urban greening to cool towns and cities: a systematic review of the empirical evidence. Landsc. Urban Plan. 97, 147e155. Brager, G.S., de Dear, R., 1998. Thermal adaptation in the built environment: a literature review. Energy Build. 27, 83e96. Ca, V.T., Asaeda, T., Abu, E.M., 1998. Reductions in air conditioning energy caused by a nearby park. Energy Build. 29, 83e92. Cadot, E., Rodwin, V.G., Spira, A., 2007. In the heat of the summer: lessons from the heat waves in Paris. J. Urban Health 84 (4), 466e468. Changnon, S.A., Kunkel, K.E., Reinke, B.C., 1996. Impacts and responses to the 1995 heat wave: a call to action. Bull. Am. Meteorol. Soc. 77, 1497e1505.


10


Coutts, A., Beringer, J., Tapper, N., 2007. Impact of increasing urban density on local climate: spatial and temporal variations in the surface energy balance in Melbourne, Australia. J. Appl. Meteorol. Climatol. 46, 477e493. CSIRO, 2007. Climate Change in Australia e Technical Report. Delsante, A.E., 2005. Is the new generation of building energy rating software up to the task? e a review of AccuRate. In: ABCB Conference ‘Building Australia’s Future 2005’, Surfers Paradise, Australia, 11e15 September 2005. DHS, January 2009. Heatwave in Victoria: an Assessment of Health Impacts. Victorian Government Department of Human Services, Victoria, 2009. http:// www.health.vic.gov.au/chiefhealthofficer/publications/heatwave.htm (accessed in January 2014). Djongyang, N.I., Tchinda, R., Njomo, D., 2010. Thermal comfort: a review paper. Renew. Sustain. Energy Rev. 14, 2626e2640. Donaldson, G.C., Keatinge, W.R., Nayha, S., 2003. Changes in summer temperature and heat related mortality since 1971 in North Carolina, South Finland, and Southeast England. Environ. Res. 91, 1e7. Epstein, Y., Moran, D., 2006. Thermal comfort and the heat stress indices. Ind. Health 44, 389e398. Haines, A., Kovats, R.S., Campbell-Lendrum, D., Corvalan, C., 2006. Climate change and human health: impacts, vulnerability, and mitigation. Lancet 367 (9528), 2101e2109. Huang, C., Barnett, A.G., Wang, X., Tong, S., 2012. Effects of extreme temperatures on years of life lost for cardiovascular deaths: a time series study in Brisbane, Australia. Circul.: Cardiovasc. Qual. Outcomes 5 (5), 609e614. Hurley, P., Physick, W., Luhar, A., 2005. TAPM e a practical approach to prognostic meteorological and air pollution modeling. Environ. Model Softw. 20, 737e752. Hwang, R.L., Lin, T.P., Cheng, M.J., Chien, J.H., 2007. Patient thermal comfort requirement for hospital environments in Taiwan. Build. Environ. 42, 2980e 2987. IPCC (Intergovernmental Panel on Climate Change), 2007. Climate Change 2007 e the Fourth Assessment Report. Cambridge University. Kalkstein, L.S., Sailor, D., Shickman, K., Sheridan, S., Vanos, J., 2014. Assessing the Health Impacts of Urban Heat Island Reduction Strategies in the District of Columbia. http://www.coolrooftoolkit.org/wp-content/uploads/2013/10/DCHeat-Mortality-Study-for-DDOE-FINAL.pdf (accessed in January 2014). Liu, K., Bass, B., 2005. Performance of Green Roof Systems. Report No. NRCC-47705. National Research Council Canada, Toronto, Canada. Loughnan, M., Nicholls, N., Tapper, N., 2010. Mortalityetemperature thresholds for ten major population centres in rural Victoria, Australia. Health & Place 16, 1287e1290. Luber, G., 2008. Climate change and extreme heat events. Am. J. Prev. Med. 35 (5), 429e435. McMichael, A.J., Woodruff, R.E., Hales, S., 2006. Climate change and human health: present and future risks. Lancet 367, 859e869.

Memon, R.A., Leung, D.Y.C., Liu, C.H., 2008. A review on the generation, determination and mitigation of Urban Heat Island. J. Environ. Sci. 20, 120e128. Morris, C.J.G., Simmonds, I., Plummer, N., 2001. Quantification of the influences of wind and cloud on the nocturnal urban heat island of a large city. J. Appl. Meteorol. 40, 169e182. Nicholls, N., Skinner, C., Loughnan, M., Tapper, N., 2008. A simple heat alert system for Melbourne, Australia. Int. J. Biometeorol. 52, 375e384. PwC report, Nov. 2011. Protecting Human Health and Safety During Severe and Extreme Heat Events. PricewaterhouseCoopers Australia (PwC) Report. Ren, Z., Chen, D., 2010. Enhanced air flow modelling for AccuRate e a nationwide house energy rating tool in Australia. Build. Environ. 45, 1276e1286. Robine, J.M., Cheung, S.L.K., Roy, S.L., Oyen, H.V., Griffiths, C., Michel, J.P., Herrmann, F.R., 2008. Death toll exceeded 70,000 in Europe during the summer of 2003. C. R. Biol. 331, 171e178. Rosenfeld, A.H., Akbari, H., Romm, J.J., Pomerantz, M., 1998. Cool communities: strategies for heat island mitigation and smog reduction. Energy Build. 28, 51e62. Rosenzweig, C., Solecki, W., Parshall, L., Gaffin, S., Lynn, B., Goldberg, R., Cox, J., Hodges, S., 2006. Mitigating New York City’s heat island with urban Forestry, living roofs, and Light surfaces. In: Sixth Symposium on the Urban Environment and Forum on Managing our Physical and Natural Resources, American Meteorological Society, January 31, 2006, Atlanta, GA. Smith, G.B., Aguilar, J.L.C., Gentle, A.R., Chen, D., 2012. Multi-parameter sensitivity analysis: a design methodology applied to energy efficiency in temperate climate houses. Energy Build. 55, 668e673. State of Australian Cities, 2013. Major Urban Unit. Department of Infrastructure and Transport, 2013. Susca, T., Gaffin, S.R., Dell’Osso, G.R., 2011. Positive effects of vegetation: urban heat island and green roofs. Environ. Pollut. 159, 2119e2126. Thatcher, M., Hurley, P., 2012. Simulating Australian urban climate in a mesoscale atmospheric numerical model. Boundary-Layer Meteorol. 142, 149e175. Tong, S., Ren, C., Becker, N., 2010. Excess deaths during the 2004 heatwave in Brisbane, Australia. J. Biometeorol. 54 (4), 393e400. Tong, S.L., Wang, X.Y., Yu, W.W., Chen, D., Wang, X.M., 2014. The impact of heatwaves on mortality in Australia: a multi-city study. Br. Med. J. Open. 4, e003579 http://dx.doi.org/10.1136/bmjopen-2013-003579. Walsh, P.J., Delsante, A., 1983. Calculation of the thermal behaviour of multi-zone buildings. Energy Build. 5, 231e242. Wong, J.K.W., Lau, L.S.K., 2013. From the ‘urban heat island’ to the ‘green island’? A preliminary investigation into the potential of retrofitting green roofs in Mongkok district of Hong Kong. Habitat Int. 39, 25e35. Xu, T., Sathaye, J., Akbari, H., Garg, V., Tetali, S., 2012. Quantifying the direct benefits of cool roofs in an urban setting: reduced cooling energy use and lowered greenhouse gas emissions. Build. Environ. 48, 1e6. Yu, C., Hien, W.N., 2006. Thermal benefits of city Parks. Energy Build. 38 (2), 105e120.


Urban vegetation and heat-related mortality in Seoul, Korea.

Modification of Heat-Related Mortality in an Elderly Urban Population by Vegetation (Urban Green) and Proximity to Water (Urban Blue): Evidence from Lisbon, Portugal.

Projected heat-related mortality in the U.S. urban northeast.

Is planned adaptation to heat reducing heat-related mortality and illness? A systematic review.

Methods to Estimate Acclimatization to Urban Heat Island Effects on Heat- and Cold-Related Mortality.

Heat-related mortality in India: excess all-cause mortality associated with the 2010 Ahmedabad heat wave.

Prevalent vegetation growth enhancement in urban environment.

Challenges associated with projecting urbanization-induced heat-related mortality.

The Two Ways of Assessing Heat-Related Mortality and Vulnerability.

Real-time surveillance of heat-related morbidity: Relation to excess mortality associated with extreme heat.

Learning to take the heat declines in U.S. heat-related mortality.

Heat-related mortality and adaptation to heat in the United States.

Urban Vegetation Cover and Vegetation Change in Accra, Ghana: Connection to Housing Quality.

Reducing neonatal mortality-reply.

Temperature-related mortality in 17 large Chinese cities: how heat and cold affect mortality in China.

Reducing perinatal mortality.

Urban and Transport Planning Related Exposures and Mortality: A Health Impact Assessment for Cities.

Towards More Comprehensive Projections of Urban Heat-Related Mortality: Estimates for New York City under Multiple Population, Adaptation, and Climate Scenarios.

Heat-related mortality in a warming climate: projections for 12 U.S. cities.

Reducing mortality from meningococcal disease.

Reducing the infant mortality rate.

Removal of atmospheric particulates by urban vegetation: implications for human and vegetative health.

Urban-hazard risk analysis: mapping of heat-related risks in the elderly in major Italian cities.

Satellite-derived subsurface urban heat island.