Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Evaluating the Impact of Urban Green Spaces and Vegetation Characteristics on Land Surface Temperature Across Swiss Cities Using Machine Learning

Schärer, Alina LU (2025) In Master Thesis in Geographic Information Science GISM01 20251
Dept of Physical Geography and Ecosystem Science
Abstract
Urban Green Spaces (UGS) help mitigate the Urban Heat Island effect by lowering Land
Surface Temperature (LST). The extent of this cooling effect is influenced by vegetation
characteristics and how green spaces are spatially arranged in different urban settings.
This thesis investigates the relationship between UGS, urban morphology, and LST in Zurich, Lausanne, and Geneva using high-resolution (10m) remote sensing and machine learning. A detailed UGS map was created from Sentinel-2 and OpenStreetMap data, and a Random Forest model was applied to each city to evaluate the influence of different variables on LST. The model for Lausanne achieved the highest R² value with 0.97, followed by Zurich with 0.94, and Geneva at 0.83.
The results... (More)
Urban Green Spaces (UGS) help mitigate the Urban Heat Island effect by lowering Land
Surface Temperature (LST). The extent of this cooling effect is influenced by vegetation
characteristics and how green spaces are spatially arranged in different urban settings.
This thesis investigates the relationship between UGS, urban morphology, and LST in Zurich, Lausanne, and Geneva using high-resolution (10m) remote sensing and machine learning. A detailed UGS map was created from Sentinel-2 and OpenStreetMap data, and a Random Forest model was applied to each city to evaluate the influence of different variables on LST. The model for Lausanne achieved the highest R² value with 0.97, followed by Zurich with 0.94, and Geneva at 0.83.
The results of this thesis show that the Normalised Difference Vegetation Index (NDVI) and the proportion of built-up area are the most important predictors in all three cities. Areas with a high NDVI are associated with lower temperatures, while areas with a high proportion of built-up areas tend to exhibit higher LST. UGS are most effective when arranged in large, continuous patches, and high vegetation provides greater cooling than low vegetation. Differences across cities suggest that local factors like topography affect variable importance. These results highlight the need to consider vegetation structure and configuration in urban heat mitigation. (Less)
Popular Abstract
Cities are often warmer than their surroundings because of dense buildings, paved
surfaces, and limited greenery. Green spaces such as parks, trees, and grass-covered areas can help lower surface temperatures. Their cooling effect depends not only on the extent of greenery across the area, but also on the type of greenery, the shape and size of green areas, and how cities are built.
This project looked at surface temperatures in Zurich, Lausanne, and Geneva to understand what makes green spaces more effective. Satellite images and openly available data were used to map greenery and built-up areas. A machine learning model helped identify which features are most closely linked to higher or lower temperatures.
The results show that green,... (More)
Cities are often warmer than their surroundings because of dense buildings, paved
surfaces, and limited greenery. Green spaces such as parks, trees, and grass-covered areas can help lower surface temperatures. Their cooling effect depends not only on the extent of greenery across the area, but also on the type of greenery, the shape and size of green areas, and how cities are built.
This project looked at surface temperatures in Zurich, Lausanne, and Geneva to understand what makes green spaces more effective. Satellite images and openly available data were used to map greenery and built-up areas. A machine learning model helped identify which features are most closely linked to higher or lower temperatures.
The results show that green, healthy vegetation is linked to cooler areas, while densely built-up areas are often much warmer. Green spaces are more effective when they are large and connected rather than small and scattered. The exact effects vary from city to city depending on urban structure and natural features such as elevation.
These insights can help city planners design more comfortable and climate-resilient urban environments, and guide smarter decisions on how and where to invest in green infrastructure. It is not just a question of adding more green space, but of placing it in ways that make a real difference. This is particularly important as cities grow and extreme heat events become more frequent due to climate change. (Less)
Please use this url to cite or link to this publication:
author
Schärer, Alina LU
supervisor
organization
course
GISM01 20251
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Geography, GIS, Urban Green Spaces, Urban Heat Island, Land Surface Temperature, NDVI, Remote Sensing, Random Forest, Landscape Configuration
publication/series
Master Thesis in Geographic Information Science
report number
194
language
English
id
9208585
date added to LUP
2025-07-31 12:29:12
date last changed
2025-07-31 12:29:12
@misc{9208585,
  abstract     = {{Urban Green Spaces (UGS) help mitigate the Urban Heat Island effect by lowering Land
Surface Temperature (LST). The extent of this cooling effect is influenced by vegetation
characteristics and how green spaces are spatially arranged in different urban settings.
This thesis investigates the relationship between UGS, urban morphology, and LST in Zurich, Lausanne, and Geneva using high-resolution (10m) remote sensing and machine learning. A detailed UGS map was created from Sentinel-2 and OpenStreetMap data, and a Random Forest model was applied to each city to evaluate the influence of different variables on LST. The model for Lausanne achieved the highest R² value with 0.97, followed by Zurich with 0.94, and Geneva at 0.83.
The results of this thesis show that the Normalised Difference Vegetation Index (NDVI) and the proportion of built-up area are the most important predictors in all three cities. Areas with a high NDVI are associated with lower temperatures, while areas with a high proportion of built-up areas tend to exhibit higher LST. UGS are most effective when arranged in large, continuous patches, and high vegetation provides greater cooling than low vegetation. Differences across cities suggest that local factors like topography affect variable importance. These results highlight the need to consider vegetation structure and configuration in urban heat mitigation.}},
  author       = {{Schärer, Alina}},
  language     = {{eng}},
  note         = {{Student Paper}},
  series       = {{Master Thesis in Geographic Information Science}},
  title        = {{Evaluating the Impact of Urban Green Spaces and Vegetation Characteristics on Land Surface Temperature Across Swiss Cities Using Machine Learning}},
  year         = {{2025}},
}