Spatial Patterns of Anthropogenic Heat and Urban Density of an Indian Metropolitan City

Abstract

Anthropogenic heat intensity arises from levels of population, buildings and vehicle densities. Population and built-up densities are very high in the cities of developing countries, which may have an impact on heat generated from metabolism and buildings differently compared to developed countries. Hence, this study investigated the magnitude of anthropogenic heat in different land uses and areas with different built-up densities pertaining to Indian metropolises; Bengaluru metropolitan area was selected for this study. The maximum metabolic heat (22.8 W/m²), vehicular heat (87.2 W/m²) and building heat (443.0 W/m²) were found in the high-density residential grids and the mixed-use grids in the city centre area during 2017. The lowest value (0.1 W/m²) was found in the low-density residential areas, public and semi-public areas, restricted areas and agricultural areas. A high positive correlation value (0.8 in 2011 and 0.72 in 2017) was found between non-residential building surface fractions and anthropogenic heat.

Keywords

Anthropogenic heat metabolic heat building heat vehicular heat building surface fraction

Introduction

Anthropogenic heat is the added source of heat to surface energy balance in urban areas (Oke, 1988). In outdoor thermal environment research, the study of anthropogenic heat on the development of urban heat islands has become crucial as the building and transport sectors contribute heat flux majorly to the outdoor thermal environment (Allen et al., 2011; Bilgen, 2014). The rapid growth of the building and transport sectors in developing countries (du Can & Price, 2008) is a matter of concern for the urban outdoor thermal environment. However, very few studies have focused on this subject in developing countries (Kotharkar et al., 2018). On the global scale, the energy released by anthropogenic activities is a small fraction of the energy of the sun intercepted by the earth, but the density of heat emission can be higher in urban areas and might contribute to the local urban heat island phenomenon (Smith et al., 2009).

The magnitude of anthropogenic heat varies from city to city and depends on population density, per capita energy use, urban transport systems, commercial and industrial activities, and the climate zone (Quah & Roth, 2012). Anthropogenic heat (Q_F) is generated mainly from three major sources—human metabolism (Q_M), vehicles (Q_V) and the building sector (Q_B)—and is represented by the equation Q_F = Q_M + Q_V + Q_B (Oke, 1988). The building and transport sectors are the major energy consumption sectors that consume more than 60% of the total energy used in urban areas globally (Bilgen, 2014). Allen et al. (2011) found that the heat released from buildings contributes to anthropogenic heat at large (89%–96%) globally. Energy use and emissions towards the environment could be affected by an increase in population density over the next several decades (Liddle, 2015).

India has the second largest population in the world at the time of this study, and by 2028 the country will surpass the present largest populated country, China (Dasvarma, 2015). However, India’s per capita emissions are lower than those of other developing countries but are rising rapidly (Babu & Kaechele, 2015). It is expected that India’s per capita emissions will continue to increase with its fast-growing economic activities. Energy demand is growing steadily with strong economic growth in different energy consumption sectors in India (Zeshan & Ahmed, 2013). Indian metropolitan cities have very high population densities, and they vary spatially by day- and night-time activities. The day- and night-time spatial variation of the population has not been considered in previous studies (Allen et al., 2011; Chow et al., 2014; Sailor & Lu, 2004; Smith et al., 2009; Sun et al., 2018) as the contribution of metabolic heat is very low in low-density cities. However, considering the density of development in Indian cities, the spatial variation of day and night time population may have a significant influence on the spatial variation of metabolic heat. In addition, the floating population during the daytime may add to metabolic heat in non-residential areas significantly.

Previous studies on anthropogenic heat (Sailor & Lu, 2004; Smith et al., 2009; Sun et al., 2018) focused mainly on the seasonal variation of anthropogenic heat and found greater emissions during winter due to the demand for building heating. Other studies (Chow et al., 2014; Quah & Roth, 2012) also focused on anthropogenic heat variation but only among the commercial, high-density residential and low-density residential land uses. However, these studies are found to be lacking in considering the various densities of urban extent, including other land uses like public and semi-public or industrial land uses. Though the previous studies focused on the relationship between anthropogenic heat and land use, only very few focused on the relationship between anthropogenic heat and building density in different land uses.

In the previous study by Quah and Roth (2012), vehicular heat was calculated based on the number of automobile vehicles by mode. Quah and Roth (2012) and Chow et al. (2014) carried out their research mainly in developed countries where the percentage of two-wheelers is significantly lower, and the volume of traffic also may be different compared to developing countries like India. In India, the share of two-wheelers is very high, around 76% in 2009 (Sharma et al., 2011). Different modes of travel and the high volume of traffic in Indian cities might change the magnitude of the vehicular heat component of anthropogenic heat compared to that of developed countries.

The main objective of this study is to investigate the magnitude of annual hourly anthropogenic heat flux for the total urban area and in micro-level grids of different land uses, such as high-density residential, low-density residential, commercial, mixed-use, industrial, public semi-public use, restricted areas and agricultural lands in different time periods. This study has taken into consideration the spatial variation of day and night-time populations in metabolic heat calculations; and also considered the floating population, which adds to daytime metabolic heat in the non-residential areas. The study also assessed the relationship between anthropogenic heat flux and the classified building surface fractions of residential and non-residential land uses, which might be useful for future sustainable urban development. Cities in developing countries lack a division-level seasonal database on energy consumption, and the metabolic and vehicular heat remains the same throughout the year (Quah & Roth, 2012). Hence, the seasonal variation of anthropogenic heat is not investigated in this study.

Study Area

The Bengaluru metropolitan area of India comprises 198 wards having a population of 8,443,675 as per Census of India, 2011. By 2021 the population was estimated to be more than 12 million. The study area covers 711.59 km² area and extends from latitude 12.83° N to 13.14° N and longitude 77.46° E to 77.78° E. Bengaluru, known as the ‘Silicon Valley of India’, is the second fastest growing city in India. The economy of the city depends on the service-based industries after the growth of Information Technology (IT)-based industries since the early 1990s (Sudhira et al., 2007). The growth rate of the built-up area is very high, and the value is around 47% during the period of five years from 2000 to 2015 (Statista Research Department, 2019). The number of motor vehicles registered in Bengaluru was found to be the second highest in India in 2017 after Delhi (Road Transport Year Book 2016–17, Ministry of Road Transport and Highways).

Methods to Estimate Anthropogenic Heat (Q_F)

Inventory-based or energy balance closure approaches are mainly used to estimate anthropogenic heat (Quah & Roth, 2012). The inventory-based approach is the most commonly used approach (Allen et al., 2011; Chow et al., 2014; Quah & Roth, 2012; Sailor & Lu, 2004; Smith et al., 2009; Sun et al., 2018), where indicators (population, vehicles and buildings) are based on simple partitioning of the sources (Grimmond, 1992). In the energy balance closure approach, anthropogenic heat is calculated as the residual value of the urban energy balance (UEB), which has an inherent problem of cumulative error (Offerle et al., 2005). Hence, this study has followed the inventory-based approach.

Research studies followed either a bottom-up model or a top-down model in an inventory-based approach to calculate anthropogenic heat (Quah & Roth, 2012). The bottom-up approach is a data-intensive and detailed method to calculate micro-level anthropogenic heat, though the aggregate city-level value may have an error. The top-down method is a less detailed approach, and the micro-level value of anthropogenic heat may have an error. Due to the difficulty in obtaining the intensive data to calculate anthropogenic heat, an approach between bottom-up and top-down models is adopted for the calculation of both urban and micro-level anthropogenic heat in this study. The top-down mechanism is used to distribute building energy, as developing countries lack an energy database at the building level. The top-down approach is more significant to achieve local-level data from regional-level data, where a relationship is established between the large-scale variable and local-scale variable (Hoffmann et al., 2016).

The total study area was subdivided into 3,055 grids measuring 500 m × 500 m (25 ha), and three major sources of anthropogenic heat, that is, metabolic heat, vehicular heat and building heat were calculated grid-wise in Watt/m² hourly. Grids were classified into high-density residential, low-density residential, commercial, mixed-use (a combination of residential and commercial land use found in India), public and semi-public, industrial, transportation and restricted area in the built-up areas; and also agricultural, vacant land, recreational, vegetation and water bodies in non-built-up areas. The land use that occupied a significant area of a grid was considered as the land use attribute of that grid. Grids were also classified based on building surface fractions of residential and non-residential grids. Finally, calculated anthropogenic heat was analysed temporally, spatially and statistically for different classes of grids in built-up areas.

Metabolic Heat (Q_M)

This component of anthropogenic heat source is due to the metabolism of human beings and animals. In the urban area, metabolic heat is contributed mainly by human beings, and hence, animals were not considered in this study. Sailor and Lu (2004) found that metabolic heat is only 2%–3% of anthropogenic heat (Q_F). In the previous studies (Allen et al., 2011; Chow et al., 2014; Sailor & Lu, 2004; Smith et al., 2009; Sun et al., 2018), the census population was distributed by distance decay function from the residential area centroid to calculate metabolic heat. The census population is the night-time residential population and does not represent the daytime population distribution. The previous studies neglected the complex mapping of day- and night-time population distribution because of their little contribution to total anthropogenic heat. However, the floating population is a significant population component observed in the non-residential areas during daytime in the metropolitan cities of developing countries, and this may have an impact on the spatial and temporal distribution of the metabolic heat component of anthropogenic heat.

The Indian census data (2011) provides the working and non-working population components. The non-working population component, mainly those staying at home, is considered the residential population both at day- and night-time. It is assumed that the working population component of one ward has flowed to non-residential areas of the same ward. Urban activities are mainly observed from 6.00 to 23.00 hrs in Bengaluru. This period was considered as the activity period, and the rest of the time was regarded as sleeping time. The human metabolism of both day- and night-times was considered. The energy produced from various human metabolisms, as listed by Fanger (1972), is considered for this study. Fanger (1972) found the daytime average metabolism rate to be 180w/m² and the night-time value as 70w/m². The following equation, adopted by Quah and Roth (2012), is used to calculate metabolic heat in this study:

Q_M = [Psⁱ × (Ms × Ts)] + [Paⁱ × (Ma × Ta)]/24 × Aⁱ

Psⁱ = Population during the night of Grid i

Paⁱ = Population during the day of Grid i

Ms = Metabolism rate while sleeping

Ma = Metabolism rate while active

Ts = Period of sleeping (23:00–6:00)

Ta = Period of activity (6:00–23:00)

Aⁱ = Grid area in m²

Grid-Wise Population Calculation

The ward-wise population was collected from the census figures (http://censusindia.gov.in). The census population comprises basically residential and night-time populations. The grid-wise night-time population (Psⁱ) was calculated by distributing the ward population within the grids based on a proportional residential building area of the respective grid. The grid-wise total daytime population (Paⁱ) is the sum of the daytime non-working ward population of residential land and the working population and floating population of non-residential land. Daytime grid-wise residential population was calculated, distributing the total non-working ward population within grids based on a proportional residential building area of the respective grid. The working population of each ward was distributed among the grids by proportional non-residential building area of the individual grid. The floating population was distributed within the non-residential building area of the central urban area. Bengaluru experiences a floating population of around 20% of the total census population (www.bangaloreone.gov.in), which is mainly the daytime population of non-residential land. The ward-wise population was projected for 2017, considering the 2001–2011 ward-wise population growth rates. The projected population for Bengaluru metropolitan area in 2017 was around 11.8 million. Grid-wise day- and night-time population densities (persons/km²) of 2011 and 2017 are shown in Figure 1.

Figure 1.

Source: Census of India, 2001, 2011 and www.bangaloreone.gov.in

Vehicular Heat (Q_V)

Top-down or bottom-up approaches were adopted by the previous studies to calculate the vehicular count. The previous studies (Allen et al., 2011; Chow et al., 2014; Smith et al., 2009) followed the bottom-up approach applied by Grimmond (1992), in which road segment-wise real-time vehicle count was considered to calculate vehicular heat with a lower error. A few researchers (Quah & Roth, 2012; Sailor & Lu, 2004; Sun et al., 2018) adopted the top-down method to calculate road segment-wise vehicle classes. The population density was the criteria for distributing vehicles in the study carried out by Sailor and Lu (2004). However, vehicle distribution is the function of road hierarchy, land use patterns, population and employment (Zhao & Chung, 2001). Quah and Roth (2012) assumed the same distribution of vehicle classes on every road segment; however, in reality, the distribution of vehicle classes is different on different types of roads. Sun et al. (2018) distributed a total vehicle number based on the road density of urban functional zones, and in this process, the vehicle type was not considered.

In this study, vehicular heat was calculated by adopting the approach applied by Grimmond (1992). The number of vehicles by type of a major road segment of the respective grid was multiplied by the length of the major road segment and the respective energy used or released (Table 1).

Table 1.

Vehicle Type-Wise Energy Used or Released.

Vehicle	Vehicle Type	Petrol/Diesel	NHC J/kg	Avg. Fuel Economy (FE) m/l *	EV = NHC/FE
Car, taxi, jeep	V1	Petrol/diesel	44.6 × 10⁶ **	12000	3.72 × 10³
Motorcycle	V2	Petrol	46.4 × 10⁶	70000	0.66 × 10³
Auto rickshaw	V3	Petrol/diesel	44.6 × 10⁶**	22000	2.03 × 10³
Bus	V4	Diesel	42.8 × 10⁶	5000	8.56 × 10³
LCV	V5	Petrol/diesel	44.6 × 10⁶**	16000	2.78 × 10³
Truck, SMV	V6	Diesel	42.8 × 10⁶	5000	8.56 × 10³

Source: International Council on Clean Transportation (2017); Goel et al. (2016); Light-Commercial Vehicles in India (2014–2015); Comprehensive Traffic and Transportation Plan for Bangalore (2011); and Quah and Roth (2012).

*In-city traffic

**Average of the heat of vehicular petrol and diesel combustion

Q_V = [∑(V_na × road length within grid_i × EV_ab)/3600]/A (Wm²)

V_n = vehicle number of a road segment in a grid i

a = vehicle type, b = Fuel type

EV = NHC_b/FE_a (energy used/released per vehicle per meter, Chow et al., 2014)

NHC = net heat of vehicular petrol and diesel combustion

For petrol, 46.4 × 10⁶ J per kg with a density of 0.75 kg per litre, and for diesel, 42.8 × 10⁶ J per kg with a density of 0.85 kg per litre (Quah & Roth, 2012).

FE = fuel economy

Estimation of Vehicle Numbers

In the cities of developing countries, there is a limitation in obtaining accurate major road-wise vehicle count data. In this study, vehicle numbers by type for all the major roads were estimated using natural neighbour interpolation of the traffic count of 50 road sections from the traffic survey data carried out by Rail India Technical and Economic Services (RITES) Ltd. for the Comprehensive Traffic and Transportation Plan (CTTP) Bangalore, 2011. The survey covers 12 hours of traffic volume at 24 mid-blocks, 16 hours of traffic volume at 16 screen line locations and 24 hours of traffic volume at 10 outer cordons. All mid-blocks cover the major and important roads of the city. Screen lines are located along with natural barriers like rivers, canals, bridges, railway lines, etc., and outer cordons are situated at the boundary of the city. All the 50 survey points have data against mode-wise (car, taxi, jeep, motorcycle, auto rickshaw, bus, light commercial vehicle (LCV), truck, and small motor vehicle (SMV) traffic volume counts. Modes of vehicles were categorized into vehicle types based on fuel economy (m/l). The hourly vehicle count was projected for each survey point based on the growth rate of vehicle type from 2011 to 2017 (V1 = 114%, V2 = 72%, V3 = 56%, V4 = 60%, V5 = 22%, V6 = 53%, and overall = 77%) as per the motor vehicle registered in Bengaluru to achieve the 2017 modal split trend for each survey point. Bengaluru metropolitan city covers around 1,600 km of major road links (2,254 numbers) consisting of 13% trunk roads (right of way [ROW] 75m, 60m, 45m, 40m, and 30m), 15% primary roads (ROW 30m, 24m), 23% secondary roads (ROW 30m, 24m, and 18m) and 48% tertiary roads (ROW 24m, 18m, 12m, and 9m). CTTP survey points were mainly located on trunk roads, primary roads and secondary roads. A natural neighbour interpolation was adopted with a 140 m distance gap (the average distance between major roads in the city centre) and considering the CTTP survey locations as input points to achieve the vehicle count for all the major road links of the entire study area. Natural neighbour interpolation is a spatial analyst tool in ArcGIS that determines the output cell values using a linear-weighted combination set of sample points (Childs, 2004). The weight assigned is a function of the distance of the output cell location from an input point. The influence on the output cell value becomes less with the increment of distance. Figure 2 shows vehicle type-wise (V1 [car, taxi, jeep], V2 [motorcycle], V3 [auto rickshaw], V4 [bus], V5 [LCV] and V6 [truck, SMV]) traffic volume interpolated from the CTTP Bangalore traffic survey.

Figure 2.

Source: Comprehensive Traffic and Transportation Plan for Bangalore, 2011 by RITES Ltd.

Building Heat (Q_B)

Electrical energy is used for providing good lighting, thermal comfort, and the use of electrical appliances and machines in buildings. A particular portion of the input of electrical energy is converted into heat energy by the electrical loads, and the heat gets dissipated from the buildings into the surrounding environment. Two methods, namely, measurements of surface energy balance and a top-down or bottom-up inventory approach, are used to estimate the anthropogenic heat from buildings in an urban area. In the surface energy balance method, residual energy is assumed as anthropogenic heat (Sailor & Lu, 2004), whereas the building component heat remains unclear. The top-down or bottom-up inventory approach needs to collect energy data for different types of buildings on different spatial and temporal scales. Therefore it has been assumed that there is no time lag between energy consumption and heat emission; hence, energy consumed by the buildings is immediately converted into sensible waste heat (Sun et al., 2018). A few studies (Sailor & Lu, 2004) adopted the top-down inventory method to calculate building-level energy consumption, where population density has been considered as a parameter to distribute energy used. In developing countries, the low-income group or economically weaker section with very high population density occupies a less built-up area and consumes less energy, therefore population density may not be the ideal indicator in the Indian context. On the other hand, the bottom-up inventory approach on an urban scale requires accurate building-level information, which is a huge task, and the aggregate value of energy used may not match the total power supply. Therefore, a combination of top-down and bottom-up inventory approaches is used for this study.

In cities of developing countries, the collection of electrical data, both in spatial resolution and temporal resolution, is a big challenge. In this study, the division-wise energy used per unit area of the residential and non-residential buildings is considered for each grid having residential and non-residential buildings, which might be a more meaningful indicator to calculate building heat as specified by Smith et al. (2009).

This approach proposes a balance between a less detailed top-down method as well as a data-intensive bottom-up method. The modified version of the equation proposed by Smith et al. (2009) was used to calculate building heat, as stated below:

Q_B = (REC_i + NREC_i)/A_i (W/m²)

REC_i = Residential building energy consumption in grid i

NREC_i = Non-Residential building energy consumption in grid i

A_i = Grid area in m²

Grid-wise energy consumption was estimated using Bangalore Electricity Supply Company Ltd. (BESCOM) data comprising 12 divisions-wise power consumption for domestic and non-domestic purposes for the years 2011 and 2017. Domestic and non-domestic energy consumption data for 2011 and 2017 were distributed among the grids having proportional residential and non-residential building areas of the respective division (Figure 3). Tables 2 and 3 show division-wise domestic and non-domestic building energy consumption per m² for 2011–2012 and 2017–2018, respectively.

Figure 3.

Source: RMP 2015 BBMP; Google Earth Pro historical imagery 2017; and Open Street Map.

Table 2.

BESCOM Division-Wise Domestic and Non-Domestic Building Energy Consumption 2011–2012.

Sl. No.	BESCOM Division	Domestic Building Energy Consumption 2011–2012 (Mega Unit)	Domestic Building Surface Area 2011 (m²)	Unit/m²	Non-Domestic Building Energy Consumption 2011–2012 (Mega Unit)	Non-Domestic Building Surface Area 2011 (m²)	Unit/m²
1.	Hebbal	345	8,078,643	42.7	434	2,788,855	155.8
2.	HSR	382	4,720,769	81.0	632	2,732,634	231.5
3.	Indiranagar	424	7,935,456	53.5	1,296	4,799,174	270.2
4.	Jayanagar	608	10,635,817	57.1	646	4,090,869	157.9
5.	Kengeri	0	2,824,106	0.0	0	1,003,637	0.0
6.	Koramangala	390	5,693,580	68.5	1,250	3,047,979	410.2
7.	Malleshwaram	272	4,833,596	56.4	289	3,139,063	92.2
8.	Peenya	215	4,257,975	50.7	730	4,496,322	162.5
9.	Rajajinagar	338	6,302,761	53.6	328	3,125,404	105.1
10.	RR Nagar	269	3,928,778	68.6	324	2,197,269	147.5
11.	Shivajinagar	372	8,066,232	46.1	404	3,317,716	121.8
12.	Vidhana Soudha	137	2,310,722	59.3	273	3,694,167	74.1

Source: Bangalore Electricity Supply Company Ltd. (BESCOM); RMP 2015 BBMP; Google Earth Pro historical imagery 2011; and Open Street Map.

Table 3.

BESCOM Division-Wise Domestic and Non-Domestic Building Energy Consumption, 2017–2018.

Sl. No.	BESCOM Division	Domestic Building Energy Consumption 2017–2018 (Mega Unit)	Domestic Building Surface Area 2017 (m²)	Unit/m²	Non-Domestic Building Energy Consumption 2017–2018 (Mega Unit)	Non-Domestic Building Surface Area 2017 (m²)	Unit/m²
1.	Hebbal	731	8,078,643	72.1	612	3,194,613	191.8
2.	HSR	707	4,720,769	92.7	726	3,000,308	241.9
3.	Indiranagar	532	7,935,456	60.6	1,154	5,184,397	222.7
4.	Jayanagar	387	10,635,817	30.1	360	4,199,628	85.9
5.	Kengeri	278	2,824,106	63.0	260	1,219,084	213.5
6.	Koramangala	187	5,693,580	22.4	287	3,503,589	82.1
7.	Malleshwaram	499	4,833,596	97.8	407	3,292,537	123.6
8.	Peenya	320	4,257,975	70.4	292	4,687,831	62.5
9.	Rajajinagar	279	6,302,761	40.5	718	3,165,475	226.9
10.	RR Nagar	542	3,928,778	115.8	531	2,252,741	235.9
11.	Shivajinagar	149	8,066,232	17.1	286	3,489,622	81.9
12.	Vidhana Soudha	654	2,310,722	267.8	1,236	3,829,302	322.7

Source: Bangalore Electricity Supply Company Ltd. (BESCOM); RMP 2015 BBMP; Google Earth Pro historical imagery 2017; and Open Street Map.

Results

The anthropogenic heat was assessed for the total urban area, city centre area and also for the different land-use grids having a built-up area or major roads. Land use categories considered are low-density residential, high-density residential, commercial, industrial, public and semi-public, mixed-use, restricted area and agricultural area.

Anthropogenic Heat in Different Land Uses

The mean annual hourly anthropogenic heat flux was found to be 49.7 W/m² in 2011 and 58.1 W/m² in 2017, with around 2.8% annual anthropogenic incremental heat rate for the Bengaluru metropolitan area. The highest anthropogenic heat flux was found in grids located in the industrial area in 2011 (457.7 W/m²) and in the high-density residential area in 2017 (453.9 W/m²). The lowest value of 0.1 W/m² both in 2011 and 2017 was found in grids located in low-density residential areas, public and semi-public areas, restricted areas and agricultural areas. The share of metabolic heat, vehicular heat and building heat of overall anthropogenic heat were 2.1 W/m² (4%), 6.1 W/m² (12%) and 41.5 W/m² (84%), respectively, in 2011, and 2.8 W/m² (5%), 10.2 W/m² (18%) and 45.1 W/m² (77%), respectively, in 2017.

The study shows different spatial distributions of metabolic, vehicular and building heat components. The spatial distribution of building heat was found to be similar to that of total anthropogenic heat, as the building heat was found to be the maximum component of total anthropogenic heat. Higher metabolic and vehicular heat was observed in the centre of the city due to population and vehicular concentration. On the other hand, higher building heat was observed in parts of the city having industrial growth during 2011 and in high-density residential and commercial grids in the city centre during 2017 (Figures 4 and 5). Higher levels of mean metabolic heat were observed in the grids of the city centre area (8.1 W/m² in 2011 and 9.7 W/m² in 2017) compared to the other parts of the study area. Maximum metabolic heat was found in the high-density residential grid in the city centre area both in 2011 and 2017 (Table 4, Figures 4 and 5). The mean vehicular heat was also found to be higher in the city centre area (18.1 W/m² in 2011 and 33.0 W/m² in 2017) than in the total of other parts of the study area.

Figure 4.

Source: The authors.

Figure 5.

Source: The authors.

Table 4.

Area or Land Use-Wise Value of Metabolic, Vehicular, Building and Total Anthropogenic Heat.

Area/Land Use		2011				2017
Area/Land Use		Metabolic Heat (W/m²)	Vehicular Heat (W/m²)	Building Heat (W/m²)	Total Q_F (W/m²)	Metabolic Heat (W/m²)	Vehicular Heat (W/m²)	Building Heat (W/m²)	Total Q_F (W/m²)
Total BBMP	Mean	2.1	6.1	41.5	49.7	2.8	10.2	45.1	58.1
	Maximum	23.9	87.2	443.0	457.7	22.8	165.1	426.5	453.9
	Minimum	0.1	0.1	0.1	0.1	0.1	0.1	0.1	0.1
City centre area	Mean	8.1	18.1	56.0	82.2	9.7	33.0	125.7	168.4
	Maximum	23.9	87.2	117.0	155.6	22.8	165.1	399.5	447.1
	Minimum	0.4	1.0	7.2	10.1	0.5	1.8	3.4	7.8
Residential low density (<20,000 persons/km²)	Mean	1.6	5.9	38.4	45.7	1.9	11.0	38.8	51.8
	Maximum	3.2	35.1	157.5	174.5	3.8	62.2	228.0	256.2
	Minimum	0.1	0.1	0.1	0.1	0.1	0.1	1.5	1.6
Residential high density (20,000 to 177,000 persons/km²)	Mean	5.9	10.3	71.9	88.1	6.3	16.3	74.4	97.1
	Maximum	23.9	53.2	203.2	227.5	22.8	100.2	426.5	453.9
	Minimum	2.7	0.2	23.3	9.2	2.9	0.1	10.8	18.5
Commercial	Mean	5.2	16.9	90.4	112.4	6.8	30.4	180.3	217.5
	Maximum	18.0	26.9	167.6	187.8	22.8	48.5	399.1	440.0
	Minimum	1.2	1.8	32.7	48.2	0.6	3.2	38.2	59.4
Industrial	Mean	2.0	6.0	102.0	110.0	3.3	13.4	81.0	95.0
	Maximum	16.5	25.5	443.0	457.7	21.4	45.2	356.2	401.6
	Minimum	0.1	0.3	2.9	3.1	0.1	0.5	5.0	7.8
Public and semi-public	Mean	2.1	8.7	41.0	51.7	2.8	20.9	63.4	81.9
	Maximum	10.5	48.0	248.5	263.7	13.6	85.7	306.5	348.3
	Minimum	0.1	0.2	0.8	0.8	0.1	0.3	0.6	0.6
Mixed use	Mean	5.6	14.3	83.5	103.4	7.6	26.5	111.6	145.0
	Maximum	18.4	87.2	185.9	209.2	22.8	165.1	422.8	447.1
	Minimum	0.9	0.4	28.8	10.7	1.2	0.7	16.9	34.6
Restricted area	Mean	1.0	4.2	23.3	28.5	1.3	13.4	28.8	38.1
	Maximum	6.2	42.0	189.0	205.1	7.1	78.7	228.0	269.7
	Minimum	0.1	0.1	0.2	0.2	0.1	0.1	0.1	0.1
Agricultural	Mean	0.28	1.6	7.2	9.1	0.9	6.1	14.8	18.3
	Maximum	4.7	20.2	97.1	102.0	11.9	36.2	100.2	105.9
	Minimum	0.1	0.1	0.1	0.1	0.1	0.1	0.1	0.1

Source: The authors.

Note: Minimum values of metabolic heat, vehicular heat and building heat are from the grid having a minimum population, major roads and buildings, respectively, to avoid vacant grids. Maximum values are marked as bold in total BBMP area as well as within different land use areas.

*Grids considered which had total anthropogenic heat ≥0.1 W/m² to avoid vacant grids.

Maximum vehicular heat was found in the mixed-use grid in the city centre area, that is, 87.2 W/m² (56%) with 5.3 W/m² (3%) metabolic heat and 63.1 W/m² (41%) building heat in 2011. In 2017, the mean vehicular heat was found to be higher in the same mixed-use grid with a higher percentage of share, that is, 165.1 W/m² (61%) with 6.7 W/m² (3%) metabolic heat and 98.5 W/m² (36%) building heat. The mixed-use grid with maximum vehicular heat in 2011 had 260,000 vehicles per hour, of which the share of two-wheelers was 53%, and the share of cars, taxi jeeps was 30%, whereas, in 2017, the mixed-use grid with maximum vehicular heat had 475,000 vehicles per hour of which share of two-wheelers was 50%, and the share of car, taxi jeep was 35%. The minimum vehicular heat found was around 0.1 W/m² recorded in the grids, with a few numbers of vehicles in all land uses.

Maximum building heat of 443.0 W/m² (97% of total anthropogenic heat) was found in the industrial area grid with 40% building surface area in 2011. The metabolic heat and vehicular heat components of that grid were found to be 1.5 W/m² (1%) and 13.2 W/m² (2%), respectively, in 2011. In 2017, the maximum building heat of 426.5 W/m² (94%) was found in the high-density residential grid with 50% building surface area. The metabolic heat and vehicular heat components of that grid were found to be 1.5 W/m² (3%) and 13.2 W/m² (3%), respectively. The minimum building heat of 0.1 W/m² was found in public and semi-public grids, restricted area grids and agricultural grids with few buildings.

The highest (or maximum) mean hourly anthropogenic heat was found in the commercial area (112.4 W/m² in 2011 and 217.5 W/m² in 2017) followed by the mixed-use area (103.4 W/m² in 2011 and 145.0 W/m² in 2017), industrial area (110.0 W/m² in 2011 and 95.0 W/m² in 2017), high-density residential area (88.1 W/m² in 2011 and 97.1 W/m² in 2017), public and semi-public area (51.7 W/m² in 2011 and 81.9 W/m² in 2017), low-density residential area (45.7 W/m² in 2011 and 51.8 W/m² in 2017), restricted area (28.5 W/m² in 2011 and 38.1 W/m² in 2017), with the lowest mean hourly anthropogenic heat in the agricultural area (9.1 W/m² in 2011 and 18.3 W/m² in 2017).

Anthropogenic Heat in Grids with Different Building Densities

It is necessary to understand the relationship between building density and anthropogenic heat as the building heat component has the largest share in anthropogenic heat. The percentage of residential and non-residential building surface areas of each grid was calculated as residential building surface fraction (RBSF) and non-residential building surface fraction (NRBSF) to identify the correlation between building densities and anthropogenic heat for 2011 and 2017. Table 5 shows the mean anthropogenic heat in different classes of RBSF and NRBSF for 2011 and 2017. A high positive Pearson’s correlation (0.809 in 2011 and 0.723 in 2017, significant at the 0.01 level) was found between NRBSF and anthropogenic heat, whereas a medium positive Pearson’s correlation (0.47 in 2011 and 0.35 in 2017, significant at the 0.01 level) was found between RBSF and anthropogenic heat (refer to Figure 6). An increment in mean anthropogenic heat was observed in all the NRBSF classes’ grids from 2011 to 2017. There is an increment in mean anthropogenic heat in <10%, 10%–20% and 20%–30% RBSF grids in 2017 compared to 2011. In the case of 30%–40% and >40% RBSF grids, there is an increase in metabolic and vehicular heat but a decrease in building heat due to lower consumption of energy in 2017 compared to 2011. In the higher building surface fraction classification, mean anthropogenic heat was much higher in the non-residential grid than a residential grid (Table 5), which implies high building density in non-residential land use generates high anthropogenic heat, which contributes to urban heat islands.

Table 5.

Mean Anthropogenic Heat in Different Classes of RBSF and NRBSF.

RBSF/NRBSF (% Building Area of Grid)	Mean Anthropogenic Heat for RBSF (W/m²) 2011	Mean Anthropogenic Heat for RBSF (W/m²) 2017	Increase from 2011 to 2017	Mean Anthropogenic Heat for NRBSF (W/m²) 2011	Mean Anthropogenic Heat for NRBSF (W/m²) 2017	Increase from 2011 to 2017
<10	32.2	38.1	5.9	39.4	46.9	7.5
10–20	59.8	68.1	8.3	101.1	120.2	9.1
20–30	80.0	90.4	10.4	133.0	160.2	27.2
30–40	93.0 (5.5 + 9.2 + 78.3)	93.0 (5.6 + 16.5 + 70.9)	Nil	168.5	195.9	27.4
>40	105.7 (6.9 + 7.2 + 91.6)	91.1 (7.2 + 11.4 + 72.5)	–	150.3	255.0	104.7

Source: The authors.

Figure 6.

Source: The authors.

Discussion

Spatial anthropogenic heat study is necessary for regional climate modelling and is significant to understand the extent and magnitude of urban heat island (UHI). It is evident that building heat is the major component of the overall anthropogenic heat, though the increment rate of vehicular heat is high among the anthropogenic heat components. Future land use and transport plans may consider this trend to address local thermal discomfort. The computed vehicular heat (minimum 0.1 W/m² to maximum 87.2 W/m²) in Bengaluru metropolitan area in 2011 is comparable with the research carried over in Manchester city in the UK by Smith et al. (2009). They found 0 W/m² (no road) to over 70 W/m² vehicular heat at major junctions of Manchester.

It was observed that the calculated minimum anthropogenic heat (48.2 W/m² in 2011 and 59.4 W/m² in 2017) was found to be high in the commercial area compared with the calculated minimum anthropogenic heat in other land use areas. Sastry et al. (2013) also found a higher urban heat island in the commercial streets of Bengaluru, which depicts the anthropogenic heat source as significantly contributing to the rise of the urban heat island. Similarly, Sun et al. (2018) also found the highest mean anthropogenic heat intensity in the commercial area, followed by industrial, public, residential, preservation (ecosystem services) and agricultural areas in Beijing. Sun et al. (2018) in their study of Beijing found a lower mean anthropogenic heat value of 14.55 W/m² and a lower highest anthropogenic heat value of 82.3 W/m² compared to Bengaluru metropolitan area, that is, 58.1 W/m² and 453.9 W/m², respectively, in 2017. The high population and building density of Bengaluru might be the reason for the higher value of anthropogenic heat. In this study, it was found that in the city centre area, the mean anthropogenic heat was 82.2 W/m² in 2011, which is comparable to the earlier findings by Quah and Roth (2012), where the mean annual hourly anthropogenic heat flux was 85 W/m² in the commercial or city centre areas of Singapore in 2008. The mean annual hourly anthropogenic heat flux in low-density (<20,000 persons per km²) residential areas, that is, 45.7 W/m² during 2011, and in high-density (20,000 to 177,000 persons per km²) residential areas, that is, 88.1 W/m² during 2011 in the Bengaluru metropolitan area was found to be higher than the anthropogenic heat found by Quah and Roth (2012) in the low-density (7,000 persons per km²) residential areas and high-density (21,000 persons per km²) residential areas of the city centre of Singapore, that is, 11 W/m² and 13 W/m² respectively. This shows that metabolic heat is also a significant contributor in a high-density residential area as well as in the city centre area (Table 4), which was insignificant in previous studies (Allen et al., 2011; Chow et al., 2014; Sailor & Lu, 2004; Smith et al., 2009; Sun et al., 2018).

Conclusion

This study has estimated 500 m × 500 m grid-wise annual hourly anthropogenic heat flux in W/m² for high-density residential, low-density residential, commercial, mixed-use, industrial and public semi-public land uses in Bengaluru. The study has adopted both top-down and bottom-up approaches during the mapping of population, vehicles and building energy consumed to calculate micro-level to urban-level anthropogenic heat. Due to the unavailability of building level energy and population data in calculating building heat and metabolic heat, the top-down method is used, which is limited in producing accurate building level heat as building surface area is only considered as a local variable in this approach, and not the floor space index (FSI), as building level height data for the entire study area was unavailable. This study has found the anthropogenic heat component of UEB to be a significant contributor to aggravating the intensity of urban heat in commercial land of the city centre area.

The study has analysed heat emissions in different land use and building densities, which may be helpful in formulating land use policies in different kinds of urban development like transport-oriented development, urban renewal development etc., from the point of view of the urban thermal environment and hence change in the quality of life. The assessed and projected local-level population, energy and vehicular emission data can be helpful in resource allocation, planning new infrastructure projects, energy conservation, formulating an emission reduction policy and addressing global warming. This study is also helpful in estimating future anthropogenic heat flux, though it depends on human activity; working from home can change the pattern of energy use of buildings and vehicular movement and hence change the pattern of anthropogenic heat flux.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

References

Allen

, Lindberg

, & Grimmond

C. S. B.

(2011). Global to city scale urban anthropogenic heat flux: Model and variability. International Journal of Climatology, 31(13), 1990–2005.

Babu

A. T. S.

, & Kaechele

(2015). Dichotomy in carbon dioxide emissions – The case of India. Climate and Development, 7(2), 165–174.

Bilgen

(2014). Structure and environmental impact of global energy consumption. Renewable and Sustainable Energy Reviews, 38, 890–902. 10.1016/j.rser.2014.07.004

Childs

(2004). Interpolating surfaces in ArcGIS spatial analyst. ArcUser, 32–35. www.esri.com

Chow

W. T. L.

, Salamanca

, Georgescu

, Mahalov

, Milne

J. M.

, & Ruddell

B. L.

(2014). A multi-method and multi-scale approach for estimating city-wide anthropogenic heat fluxes. Atmospheric Environment, 99, 64–76. 10.1016/j.atmosenv.2014.09.053

Dasvarma

G. L.

(2015). Population and environmental issues in South Asia, South Asia 2016 (13th ed., pp. 64–77). Taylor and Francis Group.

du Can

S. R.

, & Price

(2008). Sectoral trends in global energy use and greenhouse gas emissions. Energy Policy, 36(4), 1386–1403.

Goel

, Mohan

, Guttikunda

S. K.

, & Tiwari

(2016). Assessment of motor vehicle use characteristics in three Indian cities. Transportation Research Part D: Transport and Environment, 44, 254–265. 10.1016/j.trd.2015.05.006

Grimmond

C. S. B.

(1992). The suburban energy balance: Methodological considerations and results for a mid-latitude west coast city under winter and spring conditions. International Journal of Climatology, 12(5), 481–497.

10.

Hoffmann

, Schoetter

, & Schlünzen

K. H.

(2016). Statistical-dynamical downscaling of the urban heat island in Hamburg, Germany. Meteorologische Zeitschrift, 27(2), 89–109.

11.

Kotharkar

, Ramesh

, & Bagade

(2018). Urban heat island studies in South Asia: A critical review. Urban Climate, 24, 1011–1026. 10.1016/j.uclim.2017.12.006

12.

Liddle

(2015). What are the carbon emissions elasticities for income and population? Bridging STIRPAT and EKC via robust heterogeneous panel estimates. Global Environmental Change, 31, 62–73. 10.1016/j.gloenvcha.2014.10.016

13.

Offerle

, Grimmond

C. S. B.

, & Fortuniak

(2005). Heat storage and anthropogenic heat flux in relation to the energy balance of a central European city centre. International Journal of Climatology, 25(10), 1405–1419.

14.

Oke

T. R.

(1988). The urban energy balance. Progress in Physical Geography, 12(4), 471–508.

15.

Quah

A. K. L.

, & Roth

(2012). Diurnal and weekly variation of anthropogenic heat emissions in a tropical city, Singapore. Atmospheric Environment, 46, 92–103. 10.1016/j.atmosenv.2011.10.015

16.

Sailor

D. J.

, & Lu

(2004). A top-down methodology for developing diurnal and seasonal anthropogenic heating profiles for urban areas. Atmospheric Environment, 38(17), 2737–2748.

17.

Sastry

, Majumder

, Kumar

, & Kumar

D. E. V.

. (2013). Sustainable urban development: Minimising urban heat island effect and imperviousness factor (Working Paper Series No.13-03). South Asia Network of Economic Research Institutes (SANEI), p. 14.

18.

Sharma

R. D.

, Jain

, & Singh

(2011). Growth rate of motor vehicles in India - Impact of demographic and economic development. Journal of Economic and Social Studies, 1(2), 137–153.

19.

Smith

, Lindley

, & Levermore

(2009). Estimating spatial and temporal patterns of urban anthropogenic heat fluxes for UK cities: The case of Manchester. Theoretical and Applied Climatology, 98(1), 19–35.

20.

Sudhira

H. S.

, Ramachandra

T. V.

, & Subrahmanya

M. H. B.

(2007). Bangalore. Cities, 24(5), 379–390.

21.

Sun

, Wang

, & Chen

(2018). A distributed model for quantifying temporal-spatial patterns of anthropogenic heat based on energy consumption. Journal of Cleaner Production, 170, 601–609. 10.1016/j.jclepro.2017.09.153

22.

Zeshan

, & Ahmed

(2013). Energy, environment and growth nexus in South Asia. Environment, Development and Sustainability, 15(6), 1465–1475.

23.

Zhao

, & Chung

(2001). Contributing factors of annual average daily traffic in a Florida county: Exploration with geographic information system and regression models. Transportation Research Record: Journal of the Transportation Research Board, 1769(1), 113–122.

Spatial Patterns of Anthropogenic Heat and Urban Density of an Indian Metropolitan City

Abstract

Keywords

Introduction

Study Area

Methods to Estimate Anthropogenic Heat (QF)

Metabolic Heat (QM)

Grid-Wise Population Calculation

Vehicular Heat (QV)

Estimation of Vehicle Numbers

Building Heat (QB)

Results

Anthropogenic Heat in Different Land Uses

Anthropogenic Heat in Grids with Different Building Densities

Discussion

Conclusion

Footnotes

Declaration of Conflicting Interests

Funding

References

Methods to Estimate Anthropogenic Heat (Q_F)

Metabolic Heat (Q_M)

Vehicular Heat (Q_V)

Building Heat (Q_B)