M.Sc. dissertation · Banaras Hindu University · May 2025

Bengaluru: Urban Heat, Green Spaces and Urban Development

Spatiotemporal Assessment of Urban Green Space Dynamics and Their Mitigating Influence on Urban Heat Island: A Decadal Study of Bengaluru

Urban heat island index maps of Bengaluru for 2014, 2019 and 2024, expressed as land-surface-temperature z-scores from minus 4 to 3
Fig. 1 · UHI index (LST z-score), April 2014 / 2019 / 2024Click to enlarge
Degree
M.Sc. Environmental Sciences (Earth & Atmospheric Sciences)
Institution
Institute of Environment and Sustainable Development, BHU
Study area
Bengaluru
Study periods
April 2014 · 2019 · 2024

Overview

A decade of heat and green space, mapped.

In my M.Sc. dissertation at Banaras Hindu University, I studied how Bengaluru's thermal environment and urban green spaces changed over a decade. The analysis compares April observations from 2014, 2019 and 2024, using satellite imagery to examine land-surface temperature, vegetation and built-up patterns.

I wanted to look beyond how much green space a city contains and explore how its spatial arrangement relates to the distribution of heat. To do this, I combined remote sensing with landscape metrics and spatial statistical techniques, including Moran's I and a Coupling Coordination Degree Model.

The study brings together maps, spatial patterns and statistical comparisons to examine the relationship between urban growth, green-space configuration and surface temperature. Its results also point to the importance of interpreting urban green infrastructure in spatial context rather than treating all green spaces as interchangeable.

Research questions

What the study asks

  1. How did LST, NDVI and NDBI vary across Bengaluru between 2014, 2019 and 2024?
  2. How did urban green-space configuration relate to the thermal environment?
  3. What spatial clusters and hotspots emerged in the analysed indicators?
  4. How did the coupling and coordination between green-space and thermal indicators vary spatially and over time?

Methods and data

How the analysis was done

Satellite & spatial analysis

  • Landsat-based analysis
  • Land-surface temperature (LST)
  • NDVI · NDBI
  • Land-use/land-cover analysis
  • Green Space Index (GSI)

Landscape metrics

  • Patch Density (PD)
  • Largest Patch Index (LPI)
  • Landscape Shape Index (LSI)

Statistics & coordination

  • Global & Local Moran's I / LISA
  • Coupling & Coupling Coordination (CCDM)
  • AHP weighting for GSI analysis

Figures from the dissertation

Maps and analysis

Correlation matrices for land-surface temperature, NDBI, NDVI and Green Space Index for 2014, 2019 and 2024
Fig. 2 · Indicator correlations, 2014 / 2019 / 2024Click to enlarge

As annotated: LST–NDBI 0.513 (2014), 0.662 (2019), 0.613 (2024); LST–NDVI −0.165, −0.270, −0.167; LST–GSI −0.186, −0.317, −0.162; NDVI–GSI 0.507, 0.745, 0.720. Associations, not causal estimates.

Bivariate maps combining urban heat island intensity with Green Space Index classes for 2014, 2019 and 2024
Fig. 3 · UHI × GSI bivariate mapsClick to enlarge

Urban heat intensity overlaid with Green Space Index classes (high / medium / low) for each study year.

Grid-based change maps for coupling and coordination indicators across 2014–2019 and 2019–2024, with class breaks as shown
Fig. 4 · Coupling (C) and coordination (D) change mapsClick to enlarge

Change maps for the coupling (C) and coordination (D) indicators across 2014–2019 and 2019–2024, class breaks as shown in each panel.

Grid-based change maps of the Normalized Difference Built-up Index across 2014–2019 and 2019–2024, with class breaks as shown
Fig. 5 · NDBI change maps (C, D)Click to enlarge

Built-up change across the two inter-survey periods, class breaks as shown in each panel.

Bar chart of the Heat Contribution Index by land-use category, with built-up areas contributing most positively in all three periods
Fig. 6 · Heat Contribution Index by land-use categoryClick to enlarge

Built-up areas show the highest positive contribution in all three periods, while water bodies, tree cover and agriculture sit near zero or below — read qualitatively from the chart.

Findings

What the analysis examines

The study compared April LST, vegetation and built-up indicators over three time points. The spatial analyses found statistically significant positive clustering for LST, GSI and NDBI in the analysed dataset. The CCDM analysis showed that coordination between the relevant systems changed over the study periods, with stronger coordination described from 2014 to 2019 and a decline from 2019 to 2024. The spatial maps were used to compare the distribution of thermal hotspots and green spaces across the study area.

Mean land-surface-temperature magnitudes are not stated here: the abstract and results chapter report inconsistent values, and no headline warming figure is published until the underlying calculations are reconciled.

Implications

Planning with location in mind

The study highlights why urban green-space planning needs to consider location, size, distribution and configuration alongside overall green cover. Spatially concentrated heat patterns and changes in the coordination between green-space and thermal indicators suggest that planning decisions should be informed by the local context of individual urban areas.

Limitations

What this study does and doesn't show

The analysis uses selected April periods in 2014, 2019 and 2024 rather than a continuous daily time series. It analyses remotely sensed land-surface temperature, which is not the same quantity as near-surface air temperature. The analysis describes spatial associations and coordination patterns; these alone do not prove causation.

Questions about this study?

Write with context — I read everything.

← Research Get in touch →