Quantifying diurnal and seasonal variation of surface urban heat island intensity and its associated determinants across different climatic zones over Indian cities

Figures & data

Table 1. List of studies discussing SUHI across different climatic zone considering different potential drivers

N.	Study area and climatic classification	Data source/sensors and the period	SUHI drivers and technique used	Main findings	References
1	18 Asian megacities Temperate and tropical climate regions	MODIS product (2001–2003); Landsat ETM+ (2002)	The relationship of SUHI with city size, urban surface characteristic, NDVI, built-up density was analyzed Gaussian method for SUHI estimation	Temperate climate cities were experiencing significant SUHIs in summer. Tropical cities experienced intense SUHIs in the range of 5–8°C in the dry season. Magnitude and extent of SUHIs were found highly positively correlated to the population size of the cities. Importance of vegetation in some degree, urban density in the partition of sensible and latent heat fluxes in a humid environment.	(Hung et al. 2006)
2	65 cities across the United States and Canada Köppen–Geiger climate classification	MODIS products (2003–2012); Precipitation data from US PRISM climatic group and Environment Canada (2003–2012); US Census (2010) and Canada Census (2010); NCEP-NCAR reanalysis (1972–2004)	Population and precipitation Radiation, convection, evaporation, heat storage, anthropogenic heat Community Earth System Model Surface energy balance analysis Covariance analysis	Local background condition strongly affects the SUHI magnitude by increasing daytime SUHI in humid climates but decreasing in dry climates.	(Zhao et al. 2014)
3	39 cities in mainland China Köppen–Geiger climate classification	MODIS LST (2003–2013); MCD12Q1 (2010); MYD04_3 K: AOD (2003–2013)	ΔNDVI, ΔWSA, ΔAOD, CF (crop fraction), MAT (annual mean air temperature), MAP (mean yearly precipitation), P (population) Community Land Model (CLM) simulation	Daytime: positive correlation with MAP, CF, P, MAT and negative with ΔNDVI, ΔAOD and no correlation with ΔWSA. Nighttime: a positive relationship with P, CF, ΔNDVI, ΔAOD and negative with MAP, MAT, and ΔWSA. The average haze contribution to the nighttime SUHI is 0.7 ± 0.3 K for semiarid cities, which is stronger than that in the humid climate due to a stronger longwave radiative forcing of coarser aerosols.	(Cao et al. 2016)
4	332 cities of China Köppen–Geiger climate classification	China Land-Use/Cover; Data sets (CLUDs); MODIS products (2014–2016)	Different LULC metrics and vegetation classes Spearman’s rank correlation coefficient	SUHII and its relation with LULC metrics profoundly influenced by seasonal, diurnal, and climatic factors. Built-up and vegetation have the most significant effects on SUHIs.	(Yang, Huang, and Li 2017)
5	89 cities of India	LULC data from NRSC and MODIS (2011); MODIS product (2003–2014)	NDVI, Irrigation, AOD Community Land Model (CLM) simulation	More than 60% of Indian urban areas are observed to experience a daytime UCI. Agriculture and irrigation are the two dominant drivers of UHI in India. The significant presence of atmospheric aerosols over the urban areas can influence UHI.	(Kumar et al. 2017)
6	6 cities of the United States	Meteorological input (mean radiant temperature, wind speed, relative humidity, water vapor pressure, metabolic rate)	Urban Weather Generator (UWG) to simulate the UHI during the hottest week of the typical meteorological year	Green roofs have a marginal effect on UHI mitigation in high-rise urban morphology, while increases in grass and tree coverage ratios were effective, especially in hot climates.	(Kim, Gu, and Kim 2018)
7	1449 cities of China Four different climatic zones of China	MOIDS product (2009–2010); 1 km DEM from Data Center at Lanzhou; Land use and population data of 1 km resolution from the Chinese Academy of Sciences (2010) NASA nighttime light data (2010); Temperature and precipitation from Meteorological Data Website of China (2009–2010)	ΔEVI, ΔWSA, ΔAOD, population density (PD), nighttime light intensity (NL), built-up intensity (BI), urban area size (UAS), total population (TP), mean air temperature (MAT), total precipitation (P), and humidity index (HI) Spearman and partial correlation coefficient Machine learning method	A positive correlation between daytime SUHI and ΔNL and ΔBI in an arid or semiarid region. ΔEVI negatively correlated with daytime and nighttime SUHI. ΔWSA partially negatively correlated with nighttime SUHI. MAT partially positively correlated with nighttime SUHI. Few significant partial relationships exist between SUHI and UAS, TP and ΔAOD. The explanation rates during daytime were larger than during nighttime. Both the daytime and nighttime SUHII could be well determined by drivers using the machine learning method.	(Li, Wang et al. 2019c)
8	36 megacities of China China’s climatic zones	MODIS product (2014); LULC from Landsat 8 images (2014); Temperature and precipitation from climatic data collection	Five different LULC metrics, total urban area (TUA), number of the built-up patches (NBP), aridity index Stepwise multiple linear regression model	UHI intensities of 36 cities vary temporally with a sequence of summer day > summer night > winter night > winter day. UHI significantly depends on climatic zones and seasons. A lower UHI located in the smaller built-up area with dispersed distribution when compared to the larger built-up patches. The single patch with a more complex shape would mitigate the UHI intensities.	(Yue et al. 2019)
9	155 cities of China	Meteorological data (1984–2013); Population size and gross domestic production (GDP) from statistical yearbooks	Meteorological conditions, GDP level, and population size	UHI intensity strongest in summer. Cities in arid areas had higher UHI intensities than those in humid areas. Wind speed, average precipitation, and relative humidity had a significant negative correlation with UHI intensity.	(Peng et al. 2019)
10	9500 urban clusters of worlds Köppen–Geiger climate classification	MODIS product (2000–2017); Urban extent data are from Natural Earth Global Multi-resolution Terrain Elevation Data (2010)	ΔEVI Simplified urban-extent (SUE) algorithm for SUHI estimation	Cities in arid climate show distinct diurnal and seasonal patterns, with higher surface UHI during nighttime (compared to daytime) and two peaks throughout the year. The diurnal variability in surface UHI is highest for the equatorial climate zone (0.88°C) and lowest for the arid zone (0.53°C). The seasonality is highest in the snow climate zone and lowest for the equatorial climate zone. Consistent increase in the daytime surface UHI in the urban clusters of the warm temperate climate zone (0.04°C/decade) and snow climate zone (0.05°C/decade). Arid climate zones show a statistically significant increase in the nighttime surface UHI intensity (0.03°C/decade).	(Chakraborty and Lee 2019)
11	44 cities of India	MODIS product (2000–2017)	ΔEVI, ΔNL, AOD and Tropical Rainfall Measuring Mission (TRMM)	Mean daytime SUHII is positive (up to 2°C) for most cities. Most cities show a positive trend in SUHII for monsoon and post-monsoon periods, but negative for winter and summer seasons. Influence of anthropogenic forcing on SUHII is seen.	(Raj et al. 2020)

Figure 1. Spatial distribution of selected 150 Indian cities with respect to different Köppen–Geiger classification zone. The LULC map obtained from the climate change initiative (CCI) of the European Space Agency (ESA) is reported in the background

Figure 2. Schematic diagram of selection of urban and rural areas, using (a) Ahmedabad as larger patch city, (b) Chennai as coastal patch city, and (c) comparison of large Indore and small Mhow city together with different width of rural buffer

Figure 3. Mean surface urban heat island intensity (SUHII) over 150 major Indian cities in (a) summer daytime, (b) summer nighttime, (c) winter daytime, and (d) winter nighttime across different Köppen climatic zone during 2003–2018

Table 2. Mean SUHII (for all the cities) in °C, over different climatic zones during 2003–2018

Climatic zone	Summer daytime	Winter daytime	Summer nighttime	Winter nighttime
Tropical monsoon (Am)	−0.17	0.12	0.48	0.92
Tropical wet and dry (savanna) (Aw)	0.07	0.43	0.71	1.10
Hot steppe (BSh)	−0.17	0.37	1.00	1.32
Hot desert (BWh)	−0.25	−0.33	0.82	1.15
Humid subtropical (Cfa)	1.80	0.75	0.40	0.46
Dry winter humid subtropical (Cwa)	0.35	0.65	0.92	1.00
Dry winter maritime temperate (Cwb)	0.57	0.66	0.73	0.85

Figure 4. Box plot showing the variation of SUHII in different Köppen climatic zone during 2003–2018 (SD: summer daytime; SN: summer nighttime; WD: winter daytime; and WN: winter nighttime)

Table 3. Temporal trend of SUHII (for all the cities) in the different climatic zone during 2003–2018

	Summer daytime			Winter daytime			Summer nighttime			Winter nighttime
	UHI (%)	UCI (%)	Sen’s slope (°C/year)	UHI (%)	UCI (%)	Sen’s slope (°C/year)	UHI (%)	UCI (%)	Sen’s slope (°C/year)	UHI (%)	UCI (%)	Sen’s slope (°C/year)
Tropical monsoon (Am)	100	0	0.0381	100	0	0.0149	0	100	−0.0054	50	50	−0.0014
Tropical wet and dry (savanna) (Aw)	59.26	40.74	0.0028	59.26	40.74	0.0103	77.78	22.22	0.0043	90.74	9.26	0.0109
Hot steppe (BSh)	46.15	53.85	−0.0025	41.03	58.97	−0.0034	76.92	23.08	0.0082	84.62	15.38	0.0104
Hot desert (BWh)	50	50	−0.0090	25	75	−0.0044	25	75	−0.0013	25	75	−0.0026
Humid subtropical (Cfa)	100	0	0.0371	100	0	0.0337	100	0	0.0008	0	100	−0.0084
Dry winter humid subtropical (Cwa)	65.12	34.88	0.0091	48.84	51.16	0.0013	86.05	13.95	0.0077	88.37	11.63	0.0081
Dry winter maritime temperate (Cwb)	57.14	42.86	0.0083	57.14	42.86	−0.0052	71.43	28.57	0.0042	85.71	14.29	0.0066

Figure 5. Annual variation of SUHII (°C) over selected 150 cities during 2003–2018 in (a) summer daytime, (b) summer nighttime, (c) winter daytime, and (d) winter nighttime (number of cities on X-axis is as per the serial number of cities in Table S1, provided in the supplementary material)

Figure 6. Temporal trend of SUHII over 150 Indian cities during (a) summer daytime, (b) summer nighttime, (c) winter daytime, and (d) winter nighttime across different Köppen climatic zone from 2003 to 2018

Table 4. Results of stepwise multiple linear regression for all the cities in 2003–2018. Pearson’s correlation coefficients between the SUHII and the selected driving variables are reported, as well as the overall R² and p-value of stepwise multiple linear regression

	Climate	Diurnal	Seasonal	Pearson’s correlation coefficient						Overall results
				BI	ΔEVI	ΔET	ΔWSA	ΔTI	ΔNL	R^2	p-value
UHI	Aw	Day	Summer	-	−0.53***	−0.86***	-	−0.86***	-	0.87	4.68e-22 ***
			Winter	-	−0.39***	−0.82***	0.43***	−0.80***	-	0.83	2.85e-18 ***
		Night	Summer	0.45***	-	0.31**	-	0.50***	0.34**	0.51	3.68e-06 ***
			Winter	0.38***	−0.49***	-	-	0.61***	-	0.69	9.26e-13 ***
	Cwa	Day	Summer	-	−0.70***	−0.88***	0.44***	−0.79***	-	0.94	1.00e-22 ***
			Winter	-	−0.62***	−0.86***	-	−0.58***	-	0.87	1.77e-17 ***
		Night	Summer	0.55***	-	-	−0.57***	0.39***	0.33**	0.51	1.27e-05 ***
			Winter	0.62***	−0.51***	-	−0.61***	0.62***	-	0.70	1.48e-09 ***
	Cwb	Day	Summer	−0.43***	-	−0.63***	−0.70***	−0.76	-	0.90	0.0192 *
			Winter	−0.40**	−0.53**	−0.95***	0.44**	−0.94	0.33*	0.81	0.0916 *
		Night	Summer	−0.36**	-	−0.76***	-	0.31	-	0.82	0.0121 *
			Winter	0.48***	−0.87***	−0.31**	-	-	-	0.85	0.0944 *
UCI	BSh	Day	Summer	-	−0.70***	−0.93***	-	−0.52***	-	0.91	3.54e-18 ***
			Winter	-	−0.85***	−0.87***	-	−0.47***	-	0.94	7.75e-22 ***
		Night	Summer	0.42***	−0.59***	−0.34**	−0.67***	0.59***	0.33**	0.89	4.24e-14 ***
			Winter	0.52***	−0.51***	-	−0.48**	0.68***	-	0.86	3.65e-14 ***
	BWh	Day	Summer	−0.66***	−0.64**	−0.78***	0.78**	−0.66***	-	No run
			Winter	−0.34**	−0.91***	−0.87***	0.90***	−0.87***	0.43*	No run
		Night	Summer	-	−0.97***	−0.93***	-	-	-	No run
			Winter	0.60***	−0.45***	−0.44**	−0.37**	0.31**	-	No run

- Independent variables not entered into the stepwise multiple linear regression model.

*Sig. level 0.05 < p < 0.1.

**Sig. level 0.01 < p < 0.05.

***Sig. level p < 0.01.

Figure 7. Pearson’s correlation between SUHII and six drivers (BI, ΔEVI, ΔET, ΔWSA, ΔTI, and ΔNL) over (a) tropical wet and dry (Aw), (b) dry winter humid subtropical (Cwa), (c) dry winter maritime temperate (Cwb), (d) hot steppe (BSh), and (e) hot desert (BWk) climatic zone in summer daytime (SD), winter daytime (WD), summer nighttime (SN), and winter nighttime (WN)

Figure 8. SUHII and thermal inertia (TI) relationship across the different climatic zone in (a) summer daytime, (b) winter daytime, (c) summer nighttime, and (d) winter nighttime

Figure 9. SUHI and soil moisture (SM) relationship for UHI/UCI across the different climatic zone in summer daytime (SD), winter daytime (WD), summer nighttime (SN), and winter nighttime (WN)

Figure 10. Relationships of SUHII (°C) between the daytime and nighttime during summer (a) and winter (b), and between summer and winter during the daytime (c) and nighttime (d) across India’s 150 major cities averaged over the period 2003–2018

Hung, T., D. Uchihama, S. Ochi, and Y. Yasuoka. 2006. “Assessment with Satellite Data of the Urban Heat Island Effects in Asian Mega Cities.” International Journal of Applied Earth Observation and Geoinformation 8 (1): 34–48. doi:https://doi.org/10.1016/j.jag.2005.05.003.

 Web of Science ®Google Scholar

Zhao, L., X. Lee, R. B. Smith, and K. Oleson. 2014. “Strong Contributions of Local Background Climate to Urban Heat Islands.” Nature 511 (7508): 216–219. doi:https://doi.org/10.1038/nature13462.

 PubMed Web of Science ®Google Scholar

Cao, C., X. Lee, S. Liu, N. Schultz, W. Xiao, M. Zhang, and L. Zhao. 2016. “Urban Heat Islands in China Enhanced by Haze Pollution.” Nature Communications 7 (1): 1–7. doi:https://doi.org/10.1038/ncomms12509.

 Google Scholar

Yang, Q., X. Huang, and J. Li. 2017. “Assessing the Relationship between Surface Urban Heat Islands and Landscape Patterns across Climatic Zones in China.” Scientific Reports 7 (1): 1–11. doi:https://doi.org/10.1038/s41598-017-09628-w.

 PubMedGoogle Scholar

Kumar, R., V. Mishra, J. Buzan, R. Kumar, D. Shindell, and M. Huber. 2017. “Dominant Control of Agriculture and Irrigation on Urban Heat Island in India.” Scientific Reports 7 (1): 1–10. doi:https://doi.org/10.1038/s41598-017-14213-2.

 PubMedGoogle Scholar

Kim, H., D. Gu, and H. Y. Kim. 2018. “Effects of Urban Heat Island Mitigation in Various Climate Zones in the United States.” Sustainable Cities and Society 41 (January): 841–52. doi:https://doi.org/10.1016/j.scs.2018.06.021

 Google Scholar

Li, Y., L. Wang, M. Liu, G. Zhao, H. Tian, and Q. Mao. 2019c. “Associated Determinants of Surface Urban Heat Islands across 1449 Cities in China.” Advances in Meteorology 2019: 1–14. doi:https://doi.org/10.1155/2019/4892714.

 Google Scholar

Yue, W., X. Liu, Y. Zhou, and Y. Liu. 2019. “Impacts of Urban Configuration on Urban Heat Island: An Empirical Study in China Mega-Cities.” Science of the Total Environment 671: 1036–1046. doi:https://doi.org/10.1016/j.scitotenv.2019.03.421.

 Web of Science ®Google Scholar

Peng, S., Z. Feng, H. Liao, B. Huang, S. Peng, and T. Zhou. 2019. “Spatial-Temporal Pattern Of, and Driving Forces For, Urban Heat Island in China.” Ecological Indicators 96 (September 2017): 127–132. doi:https://doi.org/10.1016/j.ecolind.2018.08.059.

 Google Scholar

Chakraborty, T., and X. Lee. 2019. “A Simplified Urban-Extent Algorithm to Characterize Surface Urban Heat Islands on A Global Scale and Examine Vegetation Control on Their Spatiotemporal Variability.” International Journal of Applied Earth Observation and Geoinformation 74 (October 2018): 269–280. doi:https://doi.org/10.1016/j.jag.2018.09.015.

 Google Scholar

Raj, S., S. K. Paul, A. Chakraborty, and J. Kuttippurath. 2020. “Anthropogenic Forcing Exacerbating the Urban Heat Islands in India.” Journal of Environmental Management 257 (December 2019): 110006. doi:https://doi.org/10.1016/j.jenvman.2019.110006.

 PubMedGoogle Scholar

Data and codes availability statement

The data and codes that support the findings of this study are available on request from the corresponding author.