Malmö UniversitySustainable Digitalisation Research Centre · Department of Urban Studies

AI-driven Urban Heat and Health Analysis using Satellite Data for Climate-Resilient Planning

Land surface temperature at 10 m from Sentinel-2 (MALMÖ-LST)

Landsat measures surface temperature at about 100 m. We learn its link to Sentinel-2 reflectance and map Malmö at 10 m, for 30 dates between 2018 and 2024. Every model we tested can be put on the maps below.

Malmö University with Malmö stad.

Compare models on one date

Pick a date, then for each map a model, its loss or training, its loss cell (U-Net models), its teacher weight (U-Net student only) and its inputs. Scroll or use + and − to zoom, drag to move: all maps follow. Hover to read the temperature at the same spot in every map.

Maps side by side
Go to
Zoom
×1.0
–
– same scale on every map
UTM 33N · hover a map

Project team

This work is part of AI-driven Urban Heat and Health Analysis using Satellite Data for Climate-Resilient Planning, a project of Malmö University and Malmö stad.

Malmö University
Nosheen AbidPostdoctoral researcher, Sustainable Digitalisation Research CentreLeads the AI work: data integration and modelling
Raoof DoorshiLand surface temperature downscaling: models and maps
Helena BohmanDepartment of Urban StudiesGIS and spatial analysis
Elnaz SarkheyliDepartment of Urban StudiesUrban planning and nature-based solutions
Malmö stad
Marcus OhlströmCoordination across the city's departments
Karl McShaneDemographic analysis and data support