Evaluating the Urban Parks Cooling Extent Using Satellite

Observations: An Alternative Approach

Tesfaye Tessema

1,2

, Dale Mortimer

and Fabio Tosti

1,2

School of Computing and Engineering, University of West London, St Mary’s Road, Ealing, London W5 5RF, U.K.

The Faringdon Research Centre for NDT and Remote Sensing, University of West London, London, W5 5RF, U.K.

Tree Service, London Borough of Ealing, Perceval House, London, U.K.

Keywords: Green Infrastructure, Satellite Remote Sensing, Landsat-8, Sentinel-2, LST.

Abstract: Green infrastructure is the cooling hub of the built environment in urban settings. These interactions could

contribute towards the reduction of the rise in temperature due to urban heat island effects. It is common

practice to evaluate the cooling extent using onsite observations. Alternatively, satellite data could be a

possible source to perform the former. We analyse and evaluate the extent using the satellite observations.

Intuitively, as we go further from a park, the cooling effect will decrease but this need to be quantified. We

analyse Landsat-8 images to generate a temperature distribution in urban environment. The Land Surface

Temperature (LST) was derived from Landsat-8 and downscaled from 30 m to 10 m using the Sentinel-2

spectral indices in the Greater London area. This gives a relatively high resolution LST variation in urban

environment. A profile over a park was extracted to observe the extent of cooling from the green infrastructure

extends towards the build environment. The cooling effect varies with the park and the effect extends up to

300 m. These observations contribute towards the urban planners to maximise the cooling benefits of urban

parks to promote urban resilience and sustainability.

1 INTRODUCTION

The urbanisation and expansion of built environment

in cities are putting a lot of loads to the heat sinks such

as parks, woodlands, and street trees, collectively

called green infrastructure (GI) (Jeyachandran et al.,

2010; Tessema et al., 2023). As a result, problem such

as the Urban Heat Iceland (UHI) effect and extreme

weather conditions are observed in cities (Almeida et

al., 2021). A high temperature in cities, poses risks to

the public health (Nieuwenhuijsen, 2021). GI play a

vital role in mitigating the UHI and moderate the

thermal variations in urban settings. Quantifying the

extent of GIs in absorbing the heat and improving

thermal comfort of the environment is essential. In

this regard, in addition to the in-situ measurements of

temperature, Earth Observation (EO) satellites can be

used as a tool to investigate surface temperature

variations in urban environments (Mackey et al.,

2012). There are various studies that have evaluated

the GIs effect at different scales (Almeida et al., 2021;

J. C. Jiménez-Muñoz et al., 2014; Mackey et al.,

2012; Vieira Zezzo et al., 2023).

Microscale weather variability study is essential

in urban environment for the wellbeing of the

population live in the cities (Almeida et al., 2021;

Vieira Zezzo et al., 2023). The extreme temperature

above the normal climate range should be monitored

to reduce the impact on public health

(Nieuwenhuijsen, 2021). In addition, the data analysis

on macroscale would contribute to the urban planners

and city councils to plan reducing the extreme heat in

urban settings.

Figure 1: Location map of the study area.

168

Tessema, T., Mortimer, D. and Tosti, F.

Evaluating the Urban Parks Cooling Extent Using Satellite Observations: An Alternative Approach.

DOI: 10.5220/0012693700003696

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 10th International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2024), pages 168-172

ISBN: 978-989-758-694-1; ISSN: 2184-500X

In this study, we investigate the variability in

temperature in the urban areas and the effect of the

green areas and parks in reducing the variability. As

a show case, we focus on the Greater London area

(Figure 1). The city has a very diverse land use,

functions, and vitality. The GI cooling effects was

quantified based on hyperspectral images of optical

observation satellites. The spatial extent of cooling

effect of selected green space was analysed. Landsat-

8 and Sentinel-2 satellite data was used to analyse the

temperature variation in the city.

2 MATERIAL AND METHOD

2.1 Data

In this research, multispectral satellite imageries of

the Landsat mission by NASA and the Sentinel-2

mission by ESA that are acquired over the London

area were used. The Landsat data is available from

2013 and the Sentinel-2 mission is available from

2015, we used here the overlapping period. Landsat-

8 OLI/TIRS has a native spatial resolution of 30 m for

the eight reflective bands (B1-B7, B9), 15 m for the

panchromatic band (B8), and 100 m for the thermal

bands (B10-B11) (Storey et al., 2014). The Landsat-8

collection 1 Level-2 surface reflectance product was

used to generate spectral indices in 30 m spatial

resolution. The Sentinel-2 MSI Level-2A imagery

was used, and the mission has no thermal band. But it

provides multispectral imagery at a spatial resolution

of 10 m and with a temporal resolution of 5 days

(Drusch et al., 2012). We use the Sentinel-2 data to

analyse and optimize the Landsat-8 data to higher

spatial resolution. The percentage cloud cover over

the images was also considered when we select the

images.

2.2 Method

The heat distribution in urban areas is controlled by a

diverse land cover distribution. The interaction

between the built environment and the green spaces

determines partly the heat dynamics of urban areas.

The thermal band from Landsat-8 allow us to map the

surface thermal radiance at 100 m ground sampling

resolution. The final distribution of the thermal band

is resampled to 30 m spatial resolution by the United

States Geological Survey. LST measures the

emission of thermal radiance from the land surface

where the incoming solar energy interacts with and

heat the ground (Hulley et al., 2019). The surface

thermal radiance of the Landsat thermal band can be

used to calculate the Land Surface Temperature

(LST). LST can be computed using an empirical

relationship between TOA brightness temperatures in

a single TIR channel (Freitas et al., 2013).

LST = A







+B









(1)

Where Tb is the TOA brightness temperature in

the TIR channel, and  is the surface emissivity for

the same channel. The coefficients (A



,B



) are

determined from linear regressions of radiative

transfer simulations.

The spectral indices such as NDVI (normalised

difference vegetation index), built-up index (NDBI),

and water index (NDWI) are determined from both

Landsat-8 and Sentinel-2 data (Defries &

Townshend, 1994; Gao, 1996; Onačillová et al.,

2022). The 10 m spatial resolution of indices from

Sentinel-2 have given an opportunity to spatially

downscale the products from Landsat-8. The

downscaling is important in urban and semi-urban

areas to analyse parameters such as LST in high

resolution. In this research, we used the method

proposed by Onačillová et al., (2022) to downscale

the spatial resolution of the LST calculated from

Landsat-8.

3 RESULTS AND DISCUSSION

The hyperspectral images were used to calculate land

cover indices such as NDVI, NDBI and NDWI

(Defries & Townshend, 1994; Gao, 1996; Varshney,

2013). From these indices, surface reflectance

products are derived to determine the land surface

temperature (LST) using Landsat-8. In addition to

this, the Sentinel-2 images were used to derive similar

land cover indices with 10 m resolution. Figure 2

shows the Sentinel-2 NDVI derived from green, red,

and near-infrared band, and the LST derived from the

green, red, near-infrared, and infrared bands from the

Landsat-8 (J. C. Jiménez-Muñoz et al., 2014). We

considered images from both satellites with a

maximum cloud coverage of 20%. As a result of high

cloud cover, we selected very limited images from the

archive. The variation in NDVI in Figure 2 (a) varies

from 0 to -1 from built environment to water bodies

and 0 to +1 from built environment to vegetations.

The NDVI index contributes for dividing the land

cover based on the chlorophyl content of the

vegetations. The white part in NDVI image represents

the built environment. Figure 2 (b) shows the LST

derived from Landsat-8 with a spatial resolution of 30

Evaluating the Urban Parks Cooling Extent Using Satellite Observations: An Alternative Approach

169

m. The temperature variation shows the average

temperature over the 2023 summertime over the

Greater London area. The temperature represented in

the figure is in degree Celsius and varies from 15

in water bodies and up to 45

C in places dominated

by buildings.

Figure 2: NDVI and LST map for Greater London. (a)

NDVI using Sentinel-2 with a 10 m spatial resolution. (b)

The LST derived from Landsat-8 with a 30 m spatial

resolution. The temperature varies from 15

C in water.

To see the variation in temperature at the

microscale, we selected the Central London area as

outlined in the black rectangle in Figure 3 (b). The

LST calculated from the Landsat-8 has a spatial

resolution of 30 m and the derived temperature

variation Figure 3 (a) and (c) acquired on May 26,

2023, and September 15, 2023, respectively. The

temperature difference between these dates on

average is about 5

C. The cooling effect of parks are

significant in May as compared to September. The

temperature in the parks in May on average was 15-

C but in September the average temperature rises

to 30

C. The cooling effect of parks in September is

very much reduced and limited to water bodies and

areas with dense trees.

LSTs in Figure 3 (b) and (d) are the downscaled

version which shows relatively variable temperature.

The downscaled map shows clear demarcation

temperature in the parks, built environments, and

linear civil structures. Hyde Park in Central London,

UK, was selected based on its size and complexity.

Figure 3: LST in Central London. On 26-05-2023, (a) is

LST 30 m resolution and (b) LST downscaled using

Sentinel-2 indices. On 15-08-2023, (c) is LST 30 m

resolution and (d) LST downscaled using Sentinel-2

indices.

The LST profile was extracted across the park as

shown in Figure 4. Figure 4 (a) and (b) shows the

temperature variation in end May and September

respectively. The LST relatively smoothly vary in the

30 m resolution (the red lines) image profiles and

showed similar pattern in both dates. The downscaled

LST shows a high frequency variability in

temperature but follows a similar general trend with

the 30 m spatial resolution LST profile.

The other important aspect of downscaling of the

LST is that it enables us to identify pocket areas of

anomalously high or low LST areas. For instance, in

Figure 4a and b, the dome shaped high temperature

anomaly at a distance from 2.8 to 3 km on the profiles

is the Winter Wonderland in Hyde Park. Places such

as train stations, large shopping centres and public

centres produces high heat anomaly in the urban

settings.

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

170

4 CONCLUSIONS AND FUTURE

DEVELOPMENT

In this study, we used Landsat thermal signal and

Sentinel-2 observations to calculate the LST. The

LST generated from Landsat at 30 m resolution is

downscaled to 10 m by implementing the spectral

indices of Sentinel-1 to calibrate the Landsat LST.

We compared two dates at the beginning and ending

months of the summer 2023. The LST profile across

the park was extracted and the extent of the cooling

effect was determined.

Figure 4: LST profile on Hyde Park in Central London, UK,

the park extent is highlighted the profile plots with shaded

rectangle. (a) shows the LST in May 2023 and (b) shows

the LST in September 2023, (c) shows the profile and the

RGB image of the area, the profile is shown by the XY red

line.

The cooling effect can be extending up to 300 m

from the border of the park. In addition, our profile

analysis showed that there is a temperature variation

within the park. The limitation in the data coverage is

the percentage of the cloud cover during this period,

and the availability of Landsat images in study area.

This work would also help authorities in developing

countries to support fast growing cities and help them

to balance the urban development and the use of parks

and public spaces towards sustainable and comfort-

oriented practices.

This research will be further developed based on

geostatistical analysis of cooling effects and the

dynamics with the built environment. Validation of

the results with the in-situ weather station database

and other measurement will be done to evaluate the

accuracy of the freely available satellite data for

urban temperature monitoring. The newly launched

thermal satellites such as the UK based satellite

SatVu (SatelliteVu,2024), which has 3.5 m resolution

thermal images would be a potential for further

research endeavors.

ACKNOWLEDGEMENTS

The Authors would like to express their sincere

thanks and gratitude to the following trusts, charities,

organisations and individuals for their generosity in

supporting this project: Lord Faringdon Charitable

Trust, The Schroder Foundation, Cazenove

Charitable Trust, Ernest Cook Trust, Sir Henry

Keswick, Ian Bond, P. F. Charitable Trust, Prospect

Investment Management Limited, The Adrian Swire

Charitable Trust, The John Swire 1989 Charitable

Trust, The Sackler Trust, The Tanlaw Foundation,

and The Wyfold Charitable Trust.

REFERENCES

Almeida, C. R., Teodoro, A. C., & Gonçalves, A. (2021).

Study of the Urban Heat Island (UHI) Using Remote

Sensing Data/Techniques: A Systematic Review.

Environments, 8(10)10.3390/environments8100105

Defries, R., S., & Townshend, J., R.G. (1994). NDVI-

derived land cover classifications at a global scale.

International Journal of Remote Sensing, 15(17), 3567-

3586.

Drusch, M., Del Bello, U., Carlier, S., Colin, O., Fernandez,

V., Gascon, F., Hoersch, B., Isola, C., Laberinti, P.,

Martimort, P., Meygret, A., Spoto, F., Sy, O.,

Marchese, F., & Bargellini, P. (2012). Sentinel-2:

ESA's Optical High-Resolution Mission for GMES

Operational Services. Remote Sensing of Environment.

Freitas, S. C., Trigo, I. F., Macedo, J., Barroso, C., Silva,

R., & Perdigão, R. (2013). Land surface temperature

from multiple geostationary satellites. International

Journal of Remote Sensing, 34(9-10), 3051-3068.

Gao, B. (1996). NDWI—A normalized difference water

index for remote sensing of vegetation liquid water

from space. Remote Sensing of Environment, 58(3).

Hulley, G. C., Ghent, D., Göttsche, F. M., Guillevic, P. C.,

Mildrexler, D. J., & Coll, C. (2019). 3 - Land Surface

Temperature. In G. C. Hulley, & D. Ghent (Eds.),

Taking the Temperature of the Earth (pp. 57-127).

Elsevier.

Evaluating the Urban Parks Cooling Extent Using Satellite Observations: An Alternative Approach

171

J. C. Jiménez-Muñoz, J. A. Sobrino, D. Skoković, C.

Mattar, & J. Cristóbal. (2014). Land Surface

Temperature Retrieval Methods from Landsat-8

Thermal Infrared Sensor Data. IEEE Geoscience and

Remote Sensing Letters, 11(10), 1840-1843.

Jeyachandran, I., Burian, S. J., & Stetson, S. W. (2010).

Estimating Urban Canopy Parameters Using Synthetic

Aperture Radar Data. Journal of Applied Meteorology

and Climatology, 49(4), 732-747.

Mackey, C. W., Lee, X., & Smith, R. B. (2012). Remotely

sensing the cooling effects of city scale efforts to reduce

urban heat island. Building and Environment, 49, 348-

358. 10.1016/j.buildenv.2011.08.004

Nieuwenhuijsen, M. J. (2021). Green Infrastructure and

Health. Annual Review of Public Health, 42(1), 317-

328. 10.1146/annurev-publhealth-090419-102511

Onačillová, K., Gallay, M., Paluba, D., Péliová, A.,

Tokarčík, O., & Laubertová, D. (2022). Combining

Landsat 8 and Sentinel-2 Data in Google Earth Engine

to Derive Higher Resolution Land Surface Temperature

Maps in Urban Environment. Remote Sensing,

14(16)10.3390/rs14164076

SatelliteVu (2024). https://www.satellitevu.com. Retrieved

24 Jan 2024, from https://www.satellitevu.com

Storey, J., Choate, M., & Moe, D. (2014). Landsat 8

Thermal Infrared Sensor Geometric Characterization

and Calibration. Remote Sensing, 6(11), 11181.

10.3390/rs61111153

Tessema, T., Uzor, S., Mortimer, D., & Tosti, F. (2023).

(2023). Estimation of tree height using radar remote

sensing in urban settings: a preliminary result.

https://doi.org/10.1117/12.2684325.

Varshney, A. (2013). Improved NDBI differencing

algorithm for built-up regions change detection from

remote-sensing data: an automated approach. Remote

Sensing Letters, 4(5), 504-512.

Vieira Zezzo, L., Pereira Coltri, P., & Dubreuil, V. (2023).

Microscale models and urban heat island studies: a

systematic review. Environmental Monitoring and

Assessment, 195(11), 1284. 10.1007/s10661-023-

11906-2

GISTAM 2024 - 10th International Conference on Geographical Information Systems Theory, Applications and Management

172